{"aliases": [], "authors": ["Oscar A. Bedoya"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e83013e816eeaf295fed06c75182b06c1aec274469568d4670e8b4cffe4af000", "submitted_at": "2008-08-12T23:22:42Z", "updated_at": "2008-08-12T23:22:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/0808.1755v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:0808.1755", "identifiers": {"arxiv": "0808.1755"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/0808.1755"}], "name": "Superstring Sigma Model Computations Using the Pure Spinor Formalism", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2008-08-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2008-08-12", "url": "https://arxiv.org/abs/0808.1755"}
{"aliases": [], "authors": ["Ueli Rutishauser", "Rodney J. Douglas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a3be76f3a45f7279a9819895708c439a4014369b8bcd9da7bd70212ad3d26112", "submitted_at": "2008-09-24T22:52:05Z", "updated_at": "2008-09-24T22:52:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/0809.4296v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:0809.4296", "identifiers": {"arxiv": "0809.4296"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/0809.4296"}], "name": "State dependent computation using coupled recurrent networks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2008-09-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2008-09-24", "url": "https://arxiv.org/abs/0809.4296"}
{"aliases": [], "authors": ["Karthikeyan Lingasubramanian", "Syed M. Alam", "Sanjukta Bhanja"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "869f06d18cfbc038ec2d30c9575570966ea82a154caa8be3c7e724802f444c82", "submitted_at": "2009-06-17T19:23:28Z", "updated_at": "2009-10-27T13:57:19Z", "version": 3, "versioned_url": "https://arxiv.org/abs/0906.3282v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:0906.3282", "identifiers": {"arxiv": "0906.3282"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/0906.3282"}], "name": "Maximum Error Modeling for Fault-Tolerant Computation using Maximum a posteriori (MAP) Hypothesis", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2009-06-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2009-10-27", "url": "https://arxiv.org/abs/0906.3282"}
{"aliases": [], "authors": ["J. C. Greer", "P. Delaney", "G. Fagas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "797dab7ed4827951d5e811f0b88cc1a2624d01efbc74396745023db03aa095a5", "submitted_at": "2009-12-17T18:05:36Z", "updated_at": "2009-12-17T18:05:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/0912.3431v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:0912.3431", "identifiers": {"arxiv": "0912.3431"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/0912.3431"}], "name": "Comment on \"Electron transport through correlated molecules computed using the time-independent Wigner function: Two critical tests\"", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2009-12-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2009-12-17", "url": "https://arxiv.org/abs/0912.3431"}
{"aliases": [], "authors": ["Laszlo B. Kish", "Sunil Khatri", "Tamas Horvath"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c54ef95b46663c787d27e991c3c1b106ab690dae5842af23be6412d3614d4fba", "submitted_at": "2010-05-10T14:24:33Z", "updated_at": "2010-11-11T23:22:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/1005.1560v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1005.1560", "identifiers": {"arxiv": "1005.1560"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1005.1560"}], "name": "Computation using Noise-based Logic: Efficient String Verification over a Slow Communication Channel", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2010-05-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2010-11-11", "url": "https://arxiv.org/abs/1005.1560"}
{"aliases": [], "authors": ["Jonathan Thornburg"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e0673aecc4ea60ef5f73e38c4ce6b6a95d48d0f446c190c414f4c989f50efce1", "submitted_at": "2010-06-18T19:54:32Z", "updated_at": "2010-06-18T19:54:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1006.3788v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1006.3788", "identifiers": {"arxiv": "1006.3788"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1006.3788"}], "name": "Highly accurate and efficient self-force computations using time-domain methods: Error estimates, validation, and optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2010-06-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2010-06-18", "url": "https://arxiv.org/abs/1006.3788"}
{"aliases": [], "authors": ["Ioan Baldea", "Horst Koppel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a54bfbc37bd5d7332b6167175f671496fc33ab1e074ba394b81233d6d7ffb9f", "submitted_at": "2010-06-24T19:42:39Z", "updated_at": "2010-06-24T19:42:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1006.4865v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1006.4865", "identifiers": {"arxiv": "1006.4865"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1006.4865"}], "name": "Reply to Comment on \"Electron transport through correlated molecules computed using the time-independent Wigner function: Two critical tests''", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2010-06-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2010-06-24", "url": "https://arxiv.org/abs/1006.4865"}
{"aliases": [], "authors": ["Kevin Leung"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aadbd3133e8c80ae902e47fbf242f696d81e1dce0a781ed26b935b36bd311d89", "submitted_at": "2010-09-21T17:57:31Z", "updated_at": "2010-09-21T17:57:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1009.4162v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1009.4162", "identifiers": {"arxiv": "1009.4162"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1009.4162"}], "name": "The Surface Potential at the Air-water Interface Computed Using Density Functional Theory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2010-09-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2010-09-21", "url": "https://arxiv.org/abs/1009.4162"}
{"aliases": [], "authors": ["Erik Saule", "Erdeniz Ö. Baş", "Ümit V. Çatalyürek"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4578d3c3b0b0e18f07245b1febad33a52a55c96b501fdea1d65e3137213c78b0", "submitted_at": "2011-04-13T18:08:57Z", "updated_at": "2011-04-13T18:08:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1104.2566v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1104.2566", "identifiers": {"arxiv": "1104.2566"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1104.2566"}], "name": "Load-Balancing Spatially Located Computations using Rectangular Partitions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2011-04-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2011-04-13", "url": "https://arxiv.org/abs/1104.2566"}
{"aliases": [], "authors": ["Sokol Kosta", "Andrius Aucinas", "Pan Hui", "Richard Mortier", "Xinwen Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a210e90f8a91f7c984bba4ec082b7b1c50431630dd67110c2e1e81acd5910722", "submitted_at": "2011-05-16T21:45:54Z", "updated_at": "2011-05-16T21:45:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1105.3232v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1105.3232", "identifiers": {"arxiv": "1105.3232"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1105.3232"}], "name": "Unleashing the Power of Mobile Cloud Computing using ThinkAir", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2011-05-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2011-05-16", "url": "https://arxiv.org/abs/1105.3232"}
{"aliases": [], "authors": ["P. A. Castillo", "M. G. Arenas", "A. M. Mora", "J. L. J. Laredo", "G. Romero", "V. M Rivas", "J. J. Merelo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8516bf38c3d514a5ba981545a86e7452ecbf0eaa9f8f473ef819ce45392089d1", "submitted_at": "2011-05-25T09:13:31Z", "updated_at": "2011-05-25T09:13:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1105.4971v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1105.4971", "identifiers": {"arxiv": "1105.4971"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1105.4971"}], "name": "Distributed Evolutionary Computation using REST", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2011-05-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2011-05-25", "url": "https://arxiv.org/abs/1105.4971"}
{"aliases": [], "authors": ["Sarah Filippi", "Chris Barnes", "Julien Cornebise", "Michael P. H. Stumpf"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ed821bbb4f6303febe4f0abab97d7532c6ce0fe0eca798791d76e43c090e6c37", "submitted_at": "2011-06-30T15:54:26Z", "updated_at": "2012-10-15T14:52:40Z", "version": 4, "versioned_url": "https://arxiv.org/abs/1106.6280v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1106.6280", "identifiers": {"arxiv": "1106.6280"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1106.6280"}], "name": "On optimality of kernels for approximate Bayesian computation using sequential Monte Carlo", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2011-06-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2012-10-15", "url": "https://arxiv.org/abs/1106.6280"}
{"aliases": [], "authors": ["T. R. Gopalakrishnan Nair", "P Jayarekha"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "747dbacc3170453ac255900c150cef4606232aa96869698cb54194c99a491c56", "submitted_at": "2012-06-03T04:42:23Z", "updated_at": "2012-06-03T04:42:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1206.0419v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1206.0419", "identifiers": {"arxiv": "1206.0419"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1206.0419"}], "name": "Pre-allocation Strategies of Computational Resources in Cloud Computing using Adaptive Resonance Theory-2", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2012-06-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2012-06-03", "url": "https://arxiv.org/abs/1206.0419"}
{"aliases": [], "authors": ["D. Prangle", "M. G. B. Blum", "G. Popovic", "S. A. Sisson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e2557be7dcb980b7a2e277f98aace41c30c1821a98b0658f283f00d4f75b418c", "submitted_at": "2013-01-14T21:47:58Z", "updated_at": "2013-01-14T21:47:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1301.3166v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1301.3166", "identifiers": {"arxiv": "1301.3166"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1301.3166"}], "name": "Diagnostic tools of approximate Bayesian computation using the coverage property", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2013-01-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2013-01-14", "url": "https://arxiv.org/abs/1301.3166"}
{"aliases": [], "authors": ["Nir Friedman", "Dan Geiger", "Noam Lotner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73fcc37132774c6279a3ebb2d5371679724bf9fe107733e21bd8efd466fb8d7f", "submitted_at": "2013-01-16T15:50:10Z", "updated_at": "2013-01-16T15:50:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1301.3855v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1301.3855", "identifiers": {"arxiv": "1301.3855"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1301.3855"}], "name": "Likelihood Computations Using Value Abstractions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2013-01-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2013-01-16", "url": "https://arxiv.org/abs/1301.3855"}
{"aliases": [], "authors": ["Dave Murray-Rust", "Peter Murray-Rust"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "38f8d515d8e5ba7e0ba7fde396ebbb23f1e983e58c748b37c83e44fd9e9486c0", "submitted_at": "2013-07-11T12:35:11Z", "updated_at": "2013-07-11T12:35:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1307.3088v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1307.3088", "identifiers": {"arxiv": "1307.3088"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1307.3088"}], "name": "The Declaratron, semantic specification for scientific computation using MathML", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2013-07-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2013-07-11", "url": "https://arxiv.org/abs/1307.3088"}
{"aliases": [], "authors": ["Brian Pepper"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce881007284bbff943e7e5a823bdbef159d577c426854a8878a9d2a9c7ed0179", "submitted_at": "2013-10-14T04:40:21Z", "updated_at": "2013-12-06T19:30:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1310.3562v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1310.3562", "identifiers": {"arxiv": "1310.3562"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1310.3562"}], "name": "Adiabatic computing using 2-local Hamiltonians on a line", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2013-10-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2013-12-06", "url": "https://arxiv.org/abs/1310.3562"}
{"aliases": [], "authors": ["Vibha Rajput", "Alok Katiyar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6eabcbb79af1f7212e269d1dfc18e4e43bbdc78fcbbe7823658841e12b006894", "submitted_at": "2013-11-08T09:17:36Z", "updated_at": "2013-11-08T09:17:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1311.1907v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1311.1907", "identifiers": {"arxiv": "1311.1907"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1311.1907"}], "name": "Proactive bottleneck performance analysis in parallel computing using openMP", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2013-11-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2013-11-08", "url": "https://arxiv.org/abs/1311.1907"}
{"aliases": [], "authors": ["Dheerendra Mishra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37c102076739ba033479f05e1750acf1e65414c812c3f80eeec966fc723c8bad", "submitted_at": "2014-01-20T04:29:25Z", "updated_at": "2014-01-20T04:29:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1401.4790v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1401.4790", "identifiers": {"arxiv": "1401.4790"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1401.4790"}], "name": "Cryptanalysis of Multi-Server Authenticated Key Agreement Scheme Based on Trust Computing Using Smart Cards and Biometrics", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-01-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2014-01-20", "url": "https://arxiv.org/abs/1401.4790"}
{"aliases": [], "authors": ["Subrata Ghosh", "Krishna Aswani", "Surabhi Singh", "Satyajit Sahu", "Daisuke Fujita", "Anirban Bandyopadhyay"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "58cb0b3ee7e8f9d4f458f0ab44e7ab689f7906ec5ef0bc7370d4946876733931", "submitted_at": "2014-01-29T04:44:08Z", "updated_at": "2014-01-29T04:44:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1401.7411v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1401.7411", "identifiers": {"arxiv": "1401.7411"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1401.7411"}], "name": "Design and Construction of a Brain-Like Computer: A New Class of Frequency-Fractal Computing Using Wireless Communication in a Supramolecular Organic, Inorganic System", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-01-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2014-01-29", "url": "https://arxiv.org/abs/1401.7411"}
{"aliases": [], "authors": ["Romain Garnier", "André Barka", "Olivier Pascal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4b97b702ea1f18a0f73cafe5db027ef3466ef022dafe4fa702f0cd8028b9a2bd", "submitted_at": "2014-02-20T14:16:33Z", "updated_at": "2014-02-20T14:16:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1402.5007v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1402.5007", "identifiers": {"arxiv": "1402.5007"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1402.5007"}], "name": "Efficient periodic band diagram computation using a finite element method, Arnoldi eigensolver and sparse linear system solver", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-02-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2014-02-20", "url": "https://arxiv.org/abs/1402.5007"}
{"aliases": [], "authors": ["Ruhi Gupta"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5a27335ed40dd9f02e9a0a4bc3f0e0913a4f2add521448367871aca00e5e727", "submitted_at": "2014-03-09T09:59:32Z", "updated_at": "2014-03-09T09:59:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1403.2043v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1403.2043", "identifiers": {"arxiv": "1403.2043"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1403.2043"}], "name": "Implementation of an efficient RBAC in Cloud Computing using .NET environment", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-03-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2014-03-09", "url": "https://arxiv.org/abs/1403.2043"}
{"aliases": [], "authors": ["Mirza Cenanovic", "Peter Hansbo", "Mats G. Larson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "70fbf2668ed9d36a00726c781fe09dc237ad6be02dc8be2f56e59d447ad302d2", "submitted_at": "2014-03-14T11:25:26Z", "updated_at": "2014-03-14T11:25:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1403.3535v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1403.3535", "identifiers": {"arxiv": "1403.3535"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1403.3535"}], "name": "Minimal surface computation using a finite element method on an embedded surface", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-03-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2014-03-14", "url": "https://arxiv.org/abs/1403.3535"}
{"aliases": [], "authors": ["Daniel John Lawson", "Niall M Adams"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "385d0a84345a8d91e25bbd3e4bf2e3cf114cfff70abb9ba78371b0c6d8265cb6", "submitted_at": "2014-03-17T10:22:16Z", "updated_at": "2014-03-17T10:22:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1403.4054v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1403.4054", "identifiers": {"arxiv": "1403.4054"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1403.4054"}], "name": "A general decision framework for structuring computation using Data Directional Scaling to process massive similarity matrices", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-03-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2014-03-17", "url": "https://arxiv.org/abs/1403.4054"}
{"aliases": [], "authors": ["Moya Chen", "Doris Xin", "Damien Woods"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "777d82287ef412cd97969a4af72e49c10ebb8af4b9d03096f3c3b8e7b59a80f6", "submitted_at": "2014-05-02T22:19:53Z", "updated_at": "2014-09-05T16:27:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1405.0527v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1405.0527", "identifiers": {"arxiv": "1405.0527"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1405.0527"}], "name": "Parallel computation using active self-assembly", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-05-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2014-09-05", "url": "https://arxiv.org/abs/1405.0527"}
{"aliases": [], "authors": ["Parul Kashyap", "Rahul Singh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "edc565fa73d24a2ff4f088a5f10f0935e3db4051f2c4ef72892f7553956039ac", "submitted_at": "2014-06-18T11:15:33Z", "updated_at": "2014-06-18T11:15:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1406.4681v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1406.4681", "identifiers": {"arxiv": "1406.4681"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1406.4681"}], "name": "Crypto multi tenant: an environment of secure computing using cloud sql", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-06-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2014-06-18", "url": "https://arxiv.org/abs/1406.4681"}
{"aliases": [], "authors": ["Giang T. Nguyen", "Federico Poloni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a897b640e080f36b9014a8dd0de0b191e6d4fae373bc9370b5038c6f1e8d3cab", "submitted_at": "2014-06-27T20:09:10Z", "updated_at": "2014-11-12T21:32:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/1406.7301v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1406.7301", "identifiers": {"arxiv": "1406.7301"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1406.7301"}], "name": "Componentwise accurate fluid queue computations using doubling algorithms", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-06-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2014-11-12", "url": "https://arxiv.org/abs/1406.7301"}
{"aliases": [], "authors": ["Holger Stengel", "Jan Treibig", "Georg Hager", "Gerhard Wellein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d97dbc4cfc7f3ac67ae1a8957f3c6bcaa79971f2bc19ae3342d8a8bee60b4485", "submitted_at": "2014-10-18T21:49:45Z", "updated_at": "2015-01-17T14:07:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1410.5010v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1410.5010", "identifiers": {"arxiv": "1410.5010"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1410.5010"}], "name": "Quantifying performance bottlenecks of stencil computations using the Execution-Cache-Memory model", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-10-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-01-17", "url": "https://arxiv.org/abs/1410.5010"}
{"aliases": [], "authors": ["Jaijeet Roychowdhury"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1831461b850140ad05899452754f76de34711af61676bbb9ce511988093fe4f", "submitted_at": "2014-10-18T23:09:47Z", "updated_at": "2015-08-11T04:38:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/1410.5016v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1410.5016", "identifiers": {"arxiv": "1410.5016"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1410.5016"}], "name": "Boolean Computation Using Self-Sustaining Nonlinear Oscillators", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-10-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-08-11", "url": "https://arxiv.org/abs/1410.5016"}
{"aliases": [], "authors": ["Ioannis Emiris", "Tatjana Kalinka", "Christos Konaxis"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3678aa2497f62747d0a27e9b8da1186f9488d88ebf41eeea1232e967105ac4d5", "submitted_at": "2014-11-11T15:13:03Z", "updated_at": "2014-11-11T15:13:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1411.2846v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1411.2846", "identifiers": {"arxiv": "1411.2846"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1411.2846"}], "name": "Sparse implicitization by interpolation: Geometric computations using matrix representations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-11-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2014-11-11", "url": "https://arxiv.org/abs/1411.2846"}
{"aliases": [], "authors": ["Kuntal Roy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "31eb4c33a8a0da280eb17388d1909206315555aae925d634e93ba2a93d130c87", "submitted_at": "2014-12-06T06:15:46Z", "updated_at": "2015-01-22T20:53:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1412.2200v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1412.2200", "identifiers": {"arxiv": "1412.2200"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1412.2200"}], "name": "Response to \"Comment on \"Ultra-low-energy non-volatile straintronic computing using single multiferroic composites\"\" [Appl. Phys. Lett. 103, 173110 (2013)]", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-12-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-01-22", "url": "https://arxiv.org/abs/1412.2200"}
{"aliases": [], "authors": ["Olivier Bournez", "Johanne Cohen", "Mikaël Rabie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0c1029409e251f7125a49039aa906b7ce65cc3690dfeac91be0d289ef007b244", "submitted_at": "2014-12-08T09:56:37Z", "updated_at": "2019-02-20T11:58:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1412.2497v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1412.2497", "identifiers": {"arxiv": "1412.2497"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1412.2497"}], "name": "Homonym Population Protocols, or Providing a Small Space of Computation Using a Few Identifiers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2014-12-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-02-20", "url": "https://arxiv.org/abs/1412.2497"}
{"aliases": [], "authors": ["Balaji Subramaniam", "Wu-chun Feng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0924ab5c486b24bc90661080fabc626dd0ebfe24203e74587180fe0c1310f043", "submitted_at": "2015-01-12T17:18:17Z", "updated_at": "2015-01-12T17:18:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1501.02724v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1501.02724", "identifiers": {"arxiv": "1501.02724"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1501.02724"}], "name": "Towards Energy-Proportional Computing Using Subsystem-Level Power Management", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-01-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-01-12", "url": "https://arxiv.org/abs/1501.02724"}
{"aliases": [], "authors": ["Kuntal Roy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc7650d4cbbe01e5858c823bc1c92d3e69162d46ac6c6d9b410c48e8ff64595a", "submitted_at": "2015-01-23T03:27:26Z", "updated_at": "2015-07-09T13:37:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1501.05941v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1501.05941", "identifiers": {"arxiv": "1501.05941"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1501.05941"}], "name": "Comment on some articles regarding magnetization switching and computing using strain-mediated multiferroic composites", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-01-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-07-09", "url": "https://arxiv.org/abs/1501.05941"}
{"aliases": [], "authors": ["Stefanie Barz", "Vedran Dunjko", "Florian Schlederer", "Merritt Moore", "Elham Kashefi", "Ian A. Walmsley"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4a9f1202f4469228e562c37f5ca09bf75f61b7971064a5a2b12e56a19c2711a2", "submitted_at": "2015-01-27T10:36:41Z", "updated_at": "2015-01-27T10:36:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1501.06730v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1501.06730", "identifiers": {"arxiv": "1501.06730"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1501.06730"}], "name": "Enhanced delegated computing using coherence", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-01-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-01-27", "url": "https://arxiv.org/abs/1501.06730"}
{"aliases": [], "authors": ["François-Henry Rouet", "Xiaoye S. Li", "Pieter Ghysels", "Artem Napov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "284b1d8a91fd59b678404e204f97753d5eeaed12f4911b114e8baca226ccd60c", "submitted_at": "2015-03-18T16:01:25Z", "updated_at": "2015-06-26T18:59:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1503.05464v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1503.05464", "identifiers": {"arxiv": "1503.05464"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1503.05464"}], "name": "A distributed-memory package for dense Hierarchically Semi-Separable matrix computations using randomization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-03-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-06-26", "url": "https://arxiv.org/abs/1503.05464"}
{"aliases": [], "authors": ["Mordechai Guri", "Matan Monitz", "Yisroel Mirski", "Yuval Elovici"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6ccfb8fc7fe1a6b4e5ca9d5ee7c5fcad3c79cab98f40572f7deaa07fcbdfe257", "submitted_at": "2015-03-26T22:14:07Z", "updated_at": "2015-03-26T22:14:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1503.07919v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1503.07919", "identifiers": {"arxiv": "1503.07919"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1503.07919"}], "name": "BitWhisper: Covert Signaling Channel between Air-Gapped Computers using Thermal Manipulations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-03-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-03-26", "url": "https://arxiv.org/abs/1503.07919"}
{"aliases": [], "authors": ["Kuntal Roy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "72b30e7793cfc68703f7b4b282e39201ef3d21da4f1e2958013a5a9caecb2818", "submitted_at": "2015-04-13T17:16:45Z", "updated_at": "2015-04-13T17:16:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1504.03907v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1504.03907", "identifiers": {"arxiv": "1504.03907"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1504.03907"}], "name": "Ultra-low-energy non-volatile straintronic computing using single multiferroic composites", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-04-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-04-13", "url": "https://arxiv.org/abs/1504.03907"}
{"aliases": [], "authors": ["Hugo Gualdron", "Robson Cordeiro", "Jose Rodrigues-Jr", "Duen Chau", "Minsuk Kahng", "U Kang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1d35ecbf8487999c4ec09c5a5c59ffd80a181c2cfb8cf6e1c0502fd350bb2f63", "submitted_at": "2015-06-03T20:56:30Z", "updated_at": "2016-09-14T20:26:33Z", "version": 5, "versioned_url": "https://arxiv.org/abs/1506.01406v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1506.01406", "identifiers": {"arxiv": "1506.01406"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1506.01406"}], "name": "M-Flash: Fast Billion-scale Graph Computation Using a Bimodal Block Processing Model", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-06-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-09-14", "url": "https://arxiv.org/abs/1506.01406"}
{"aliases": [], "authors": ["Xavier Lagorce", "Ryad Benosman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf3e3a817b7a19c62960e0be0bc12e08d49577871cb53bd223922760786c4a96", "submitted_at": "2015-07-22T15:09:07Z", "updated_at": "2015-07-22T15:09:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1507.06222v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1507.06222", "identifiers": {"arxiv": "1507.06222"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1507.06222"}], "name": "STICK: Spike Time Interval Computational Kernel, A Framework for General Purpose Computation using Neurons, Precise Timing, Delays, and Synchrony", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-07-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-07-22", "url": "https://arxiv.org/abs/1507.06222"}
{"aliases": [], "authors": ["Changsoo Je", "Min Tang", "Youngeun Lee", "Minkyoung Lee", "Young J. Kim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e846aba6270a7f2cfdb29e9cdba2a015282e8ba0a16077c036bb4c6b32843fa", "submitted_at": "2015-08-25T15:01:47Z", "updated_at": "2015-08-25T15:01:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1508.06181v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1508.06181", "identifiers": {"arxiv": "1508.06181"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1508.06181"}], "name": "PolyDepth: Real-time Penetration Depth Computation using Iterative Contact-Space Projection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-08-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-08-25", "url": "https://arxiv.org/abs/1508.06181"}
{"aliases": [], "authors": ["Sajjad AbdollahRamezani", "Kamalodin Arik", "Amin Khavasi", "Zahra Kavehvash"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e0f9ff80e2dcd6fea47d8fb44295d880d6a201ef49cd092b01890f43b4e97755", "submitted_at": "2015-09-02T19:54:03Z", "updated_at": "2015-09-02T19:54:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1509.01226v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1509.01226", "identifiers": {"arxiv": "1509.01226"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1509.01226"}], "name": "Analog Computing Using Graphene-based Metalines", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-09-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2015-09-02", "url": "https://arxiv.org/abs/1509.01226"}
{"aliases": [], "authors": ["Fereydoun Farrahi Moghaddam", "Mohamed Cheriet"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4b4cebcc4e12696ff0646e256ebece523ebc2d6a571faa640f8cf139b40131ca", "submitted_at": "2015-10-17T23:33:20Z", "updated_at": "2017-11-01T12:55:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1510.05182v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1510.05182", "identifiers": {"arxiv": "1510.05182"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1510.05182"}], "name": "Sustainability-Aware Cloud Computing Using Virtual Carbon Tax", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-10-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-11-01", "url": "https://arxiv.org/abs/1510.05182"}
{"aliases": [], "authors": ["Anna Rohrbach", "Marcus Rohrbach", "Ronghang Hu", "Trevor Darrell", "Bernt Schiele"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c9dbdd34a096069b2e6491bb3836d3b244f58516a28f0f69369c82689a20a09d", "submitted_at": "2015-11-12T01:13:47Z", "updated_at": "2017-02-17T21:02:05Z", "version": 4, "versioned_url": "https://arxiv.org/abs/1511.03745v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:1511.03745", "identifiers": {"arxiv": "1511.03745"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1511.03745"}], "name": "Grounding of Textual Phrases in Images by Reconstruction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-11-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-02-17", "url": "https://arxiv.org/abs/1511.03745"}
{"aliases": [], "authors": ["Omar Darwish", "Kurt Mehlhorn"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "28af23771f3e341dd126dc42d87be1812607fc29f364614af9bd88d8d8484f47", "submitted_at": "2015-12-18T14:41:01Z", "updated_at": "2016-05-13T11:34:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1512.05974v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1512.05974", "identifiers": {"arxiv": "1512.05974"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1512.05974"}], "name": "Improved Balanced Flow Computation Using Parametric Flow", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2015-12-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-05-13", "url": "https://arxiv.org/abs/1512.05974"}
{"aliases": [], "authors": ["Rodrigo Nogueira", "Kyunghyun Cho"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67f8697c21dabc2c4781eba46bfb60671506b413471efcdb28864c11649b3a2c", "submitted_at": "2016-02-06T14:53:02Z", "updated_at": "2016-05-20T16:26:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1602.02261v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:46Z", "role": "primary", "sha256": "5f99bcea7c1585129af46fa5c978002778dedc3d98e3a907e92030f4a5588f00", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/1602.02261v2"}], "id": "arxiv:1602.02261", "identifiers": {"arxiv": "1602.02261"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:46Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/1602.02261"}], "name": "End-to-End Goal-Driven Web Navigation", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2016-02-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebNav for converting websites into goal-directed hyperlink navigation tasks and WikiNav for evaluating language-conditioned navigation and question answering on Wikipedia.", "topics": ["web-navigation", "evaluation", "planning"], "updated_at": "2016-05-20", "url": "https://arxiv.org/abs/1602.02261"}
{"aliases": [], "authors": ["Evgenia Christoforou", "Antonio Fernández Anta", "Chryssis Georgiou", "Miguel A. Mosteiro"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db570f4f7c1af30eb3738cbda9cd694eb12e562744716b5495247cd98aeb26b0", "submitted_at": "2016-03-14T19:07:53Z", "updated_at": "2016-03-30T09:09:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1603.04394v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1603.04394", "identifiers": {"arxiv": "1603.04394"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1603.04394"}], "name": "Internet Computing: Using Reputation to Select Workers from a Pool", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-03-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-03-30", "url": "https://arxiv.org/abs/1603.04394"}
{"aliases": [], "authors": ["Giang T. Nguyen", "Federico Poloni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "24ba6eff291b83333cf461f9c6b7a5648f8575210489d3612100b36a4b169d46", "submitted_at": "2016-05-05T03:29:07Z", "updated_at": "2016-05-05T03:29:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1605.01482v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1605.01482", "identifiers": {"arxiv": "1605.01482"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1605.01482"}], "name": "Componentwise accurate Brownian motion computations using Cyclic Reduction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-05-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-05-05", "url": "https://arxiv.org/abs/1605.01482"}
{"aliases": [], "authors": ["W. H. Ball", "B. Beeck", "R. H. Cameron", "L. Gizon"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6ace750a18cb51d37771d1df851fe6f346a93697b71811e73603b3b0802337ae", "submitted_at": "2016-06-08T20:00:00Z", "updated_at": "2016-06-08T20:00:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1606.02713v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1606.02713", "identifiers": {"arxiv": "1606.02713"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1606.02713"}], "name": "MESA meets MURaM: Surface effects in main-sequence solar-like oscillators computed using three-dimensional radiation hydrodynamics simulations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-06-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-06-08", "url": "https://arxiv.org/abs/1606.02713"}
{"aliases": [], "authors": ["Tim Moon", "Jack Poulson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "04ac1fad1f009911c7976c837b8e87fe2a23aa59a00a99f3706dfb8eb7b1ebe1", "submitted_at": "2016-07-06T04:19:04Z", "updated_at": "2016-07-30T16:52:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1607.01477v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1607.01477", "identifiers": {"arxiv": "1607.01477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1607.01477"}], "name": "Accelerating eigenvector and pseudospectra computation using blocked multi-shift triangular solves", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-07-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-07-30", "url": "https://arxiv.org/abs/1607.01477"}
{"aliases": [], "authors": ["Glenn Volkema", "Gaurav Khanna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a6b72ea7c5394d9a6f4c35c290dceae328e49a31913b4a3ac58df7a06d7a19c5", "submitted_at": "2016-07-19T12:14:36Z", "updated_at": "2016-07-19T12:14:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1607.05537v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1607.05537", "identifiers": {"arxiv": "1607.05537"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1607.05537"}], "name": "Scientific Computing Using Consumer Video-Gaming Hardware Devices", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-07-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-07-19", "url": "https://arxiv.org/abs/1607.05537"}
{"aliases": [], "authors": ["Jim Griffin", "Fabrizio Leisen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "160a7d339aacd4dc92342b8d7e7c519c1f4c8d9b549ee167662f8894f1f30495", "submitted_at": "2016-08-02T15:47:42Z", "updated_at": "2017-08-31T10:21:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/1608.00874v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1608.00874", "identifiers": {"arxiv": "1608.00874"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1608.00874"}], "name": "Modelling and computation using NCoRM mixtures for density regression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-08-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-08-31", "url": "https://arxiv.org/abs/1608.00874"}
{"aliases": [], "authors": ["Philip Korman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f11f846799031f0bf9fe6784b04cdd1b2e82104e2d2a825d57c85965859fd20a", "submitted_at": "2016-08-18T19:39:22Z", "updated_at": "2016-08-18T19:39:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1608.05385v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1608.05385", "identifiers": {"arxiv": "1608.05385"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1608.05385"}], "name": "Regularization of radial solutions of $p$-Laplace equations, and computations using infinite series", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-08-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-08-18", "url": "https://arxiv.org/abs/1608.05385"}
{"aliases": [], "authors": ["Yajue Yang", "Yuanqing Wu", "Jia Pan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "709757eaaf55b9e76dab9732dd4a578e5b16fb66645cb886be8fca725cac5970", "submitted_at": "2016-09-15T02:11:16Z", "updated_at": "2016-09-15T02:11:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1609.04493v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1609.04493", "identifiers": {"arxiv": "1609.04493"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1609.04493"}], "name": "Parallel Dynamics Computation using Prefix Sum Operations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-09-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-09-15", "url": "https://arxiv.org/abs/1609.04493"}
{"aliases": [], "authors": ["Anders Pors", "Michael G. Nielsen", "Sergey I. Bozhevolnyi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "805f44633edc24170ba5b5e735cdd12a0e9e877d377d978e0b289e6e2cfebb60", "submitted_at": "2016-09-15T14:35:11Z", "updated_at": "2016-09-15T14:35:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1609.04672v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1609.04672", "identifiers": {"arxiv": "1609.04672"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1609.04672"}], "name": "Analog computing using reflective plasmonic metasurfaces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-09-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-09-15", "url": "https://arxiv.org/abs/1609.04672"}
{"aliases": [], "authors": ["Abelino Jimenez", "Bhiksha Raj"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "28e003f11dad4a1d2be839fb4d1c7736cd64a812d124fda63e7ba8dfd0da71cb", "submitted_at": "2016-09-16T19:02:04Z", "updated_at": "2016-09-23T16:12:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1609.05178v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1609.05178", "identifiers": {"arxiv": "1609.05178"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1609.05178"}], "name": "Privacy Preserving Distance Computation using Somewhat-trusted Third Parties", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-09-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2016-09-23", "url": "https://arxiv.org/abs/1609.05178"}
{"aliases": [], "authors": ["Seyed-Mohsen Moosavi-Dezfooli", "Alhussein Fawzi", "Omar Fawzi", "Pascal Frossard"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e7ad7490d7f88202f585aad7011c2b1a4a995fee2ad59c9ad3d5da22f40f4ae", "submitted_at": "2016-10-26T16:30:45Z", "updated_at": "2017-03-09T17:01:25Z", "version": 3, "versioned_url": "https://arxiv.org/abs/1610.08401v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:1610.08401", "identifiers": {"arxiv": "1610.08401"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1610.08401"}], "name": "Universal adversarial perturbations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-10-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-03-09", "url": "https://arxiv.org/abs/1610.08401"}
{"aliases": [], "authors": ["Davide Falanga", "Elias Mueggler", "Matthias Faessler", "Davide Scaramuzza"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "489803f9aee3b5f1eacc3cd8ced93c5bc0f7237769411d7c949a4844c796a28e", "submitted_at": "2016-12-01T15:03:07Z", "updated_at": "2018-04-05T14:40:20Z", "version": 6, "versioned_url": "https://arxiv.org/abs/1612.00291v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1612.00291", "identifiers": {"arxiv": "1612.00291"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1612.00291"}], "name": "Aggressive Quadrotor Flight through Narrow Gaps with Onboard Sensing and Computing using Active Vision", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2016-12-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-04-05", "url": "https://arxiv.org/abs/1612.00291"}
{"aliases": [], "authors": ["M. Omer Mushtaq", "Furrakh Shahzad", "M. Owais Tariq", "Mahina Riaz", "Bushra Majeed"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03984b6d6d0ecb170bae9497480056838fb4ef6873650aaab4f33de44ecac697", "submitted_at": "2017-02-23T09:06:07Z", "updated_at": "2017-02-23T09:06:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1702.07140v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1702.07140", "identifiers": {"arxiv": "1702.07140"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1702.07140"}], "name": "An Efficient Framework for Information Security in Cloud Computing Using Auditing Algorithm Shell (AAS)", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2017-02-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-02-23", "url": "https://arxiv.org/abs/1702.07140"}
{"aliases": [], "authors": ["Adam Strzebonski"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3fb0d1da953ceccefac54b5826652a966ede2517ac4bc4251be7f7872edb3f1d", "submitted_at": "2017-04-22T23:26:57Z", "updated_at": "2017-04-22T23:26:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1704.06856v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1704.06856", "identifiers": {"arxiv": "1704.06856"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1704.06856"}], "name": "CAD Adjacency Computation Using Validated Numerics", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2017-04-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-04-22", "url": "https://arxiv.org/abs/1704.06856"}
{"aliases": [], "authors": ["Jarilyn M. Hernández Jiménez", "Jeffrey A. Nichols", "Katerina Goseva-Popstojanova", "Stacy Prowell", "Robert A. Bridges"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3ea790044a82612a87c28fec60f6479c0e5aaca8ad5db1ccc8c8d3dfa6535e6b", "submitted_at": "2017-05-04T19:09:12Z", "updated_at": "2017-05-04T19:09:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1705.01977v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1705.01977", "identifiers": {"arxiv": "1705.01977"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1705.01977"}], "name": "Malware Detection on General-Purpose Computers Using Power Consumption Monitoring: A Proof of Concept and Case Study", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2017-05-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-05-04", "url": "https://arxiv.org/abs/1705.01977"}
{"aliases": [], "authors": ["Xinru Yan", "Ted Pedersen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "241ae1f89d67ac5f46ddb591ad8225f943afffc8aa4a56e68988b471c2778a7f", "submitted_at": "2017-05-29T16:20:21Z", "updated_at": "2017-05-29T16:20:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1705.10272v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1705.10272", "identifiers": {"arxiv": "1705.10272"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1705.10272"}], "name": "Who's to say what's funny? A computer using Language Models and Deep Learning, That's Who!", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2017-05-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-05-29", "url": "https://arxiv.org/abs/1705.10272"}
{"aliases": [], "authors": ["Max Sagebaum", "Tim Albring", "Nicolas R. Gauger"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7adf8333f9008f9994ce0383307c7169bb27dca85476c49d0e3ffaec2a151ff4", "submitted_at": "2017-09-21T09:28:36Z", "updated_at": "2017-09-21T09:28:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1709.07229v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:09Z", "role": "primary", "sha256": "a3ebf2fcfa181413d078dc3408f17bfabfac87f9aee1c78994840197df701e9d", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1709.07229", "identifiers": {"arxiv": "1709.07229"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1709.07229"}], "name": "High-Performance Derivative Computations using CoDiPack", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2017-09-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-09-21", "url": "https://arxiv.org/abs/1709.07229"}
{"aliases": [], "authors": ["Takuma Yamaguchi", "Kohei Fujita", "Tsuyoshi Ichimura", "Muneo Hori", "Maddegedara Lalith", "Kengo Nakajima"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "046ac5d6c891f6ef487116662feae1ea995fab7aff8d207ad0f6784aa87fa1c9", "submitted_at": "2017-10-24T09:51:31Z", "updated_at": "2017-10-24T09:51:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1710.08679v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1710.08679", "identifiers": {"arxiv": "1710.08679"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1710.08679"}], "name": "Implicit Low-Order Unstructured Finite-Element Multiple Simulation Enhanced by Dense Computation using OpenACC", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2017-10-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-10-24", "url": "https://arxiv.org/abs/1710.08679"}
{"aliases": [], "authors": ["Kangrui Wang", "Dalia Chakrabarty"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1525756ad7365e26aac79528879a66dc60e1ebe66688180841b547328925989f", "submitted_at": "2017-10-31T01:48:23Z", "updated_at": "2018-11-06T02:52:35Z", "version": 3, "versioned_url": "https://arxiv.org/abs/1710.11292v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1710.11292", "identifiers": {"arxiv": "1710.11292"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1710.11292"}], "name": "Correlation between Multivariate Datasets, from Inter-Graph Distance computed using Graphical Models Learnt With Uncertainties", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2017-10-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-11-06", "url": "https://arxiv.org/abs/1710.11292"}
{"aliases": [], "authors": ["Samiya Khan", "Kashish Ara Shakil", "Mansaf Alam"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8911134e61280b9a73da7d30b123726cfc1b7ec200987921ba4f75c81efea3d", "submitted_at": "2017-11-24T07:26:51Z", "updated_at": "2017-11-24T07:26:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1712.05233v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1712.05233", "identifiers": {"arxiv": "1712.05233"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1712.05233"}], "name": "Big Data Computing Using Cloud-Based Technologies, Challenges and Future Perspectives", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2017-11-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2017-11-24", "url": "https://arxiv.org/abs/1712.05233"}
{"aliases": [], "authors": ["Kei Hasegawa", "Nobuaki Fuji", "Kensuke Konishi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3137cba4546885424280fbdd22be9aa57203d16706f778adf3af75645baa9201", "submitted_at": "2018-01-09T08:02:03Z", "updated_at": "2019-02-15T03:09:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1801.02829v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1801.02829", "identifiers": {"arxiv": "1801.02829"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1801.02829"}], "name": "Improvement of accuracy of the spectral element method for elastic wave computation using modified numerical integration operators", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-01-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-02-15", "url": "https://arxiv.org/abs/1801.02829"}
{"aliases": [], "authors": ["Ivan Carnimeo", "Stefano Baroni", "Paolo Giannozzi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e75c618853743c075e2067a0df7696a8ecb3e694383ef192787994c9e095a670", "submitted_at": "2018-01-28T18:33:53Z", "updated_at": "2018-12-11T11:23:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/1801.09263v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1801.09263", "identifiers": {"arxiv": "1801.09263"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1801.09263"}], "name": "Fast hybrid density-functional computations using plane-wave basis sets", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-01-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-12-11", "url": "https://arxiv.org/abs/1801.09263"}
{"aliases": [], "authors": ["Shlomi Dolev", "Peeyush Gupta", "Yin Li", "Sharad Mehrotra", "Shantanu Sharma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f69ba53a4df9a9de0e2f9fbc283505259237c448f2e6bc89ff664431273156f5", "submitted_at": "2018-01-31T07:02:10Z", "updated_at": "2019-08-06T03:15:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/1801.10323v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1801.10323", "identifiers": {"arxiv": "1801.10323"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1801.10323"}], "name": "Privacy-Preserving Secret Shared Computations using MapReduce", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-01-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-08-06", "url": "https://arxiv.org/abs/1801.10323"}
{"aliases": [], "authors": ["Rawad Bitar", "Parimal Parag", "Salim El Rouayheb"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "46c74dbfa0944752c12e5606038ccd0e3cfe0930ccc5c1d000b5a56831a73fca", "submitted_at": "2018-02-07T21:23:44Z", "updated_at": "2018-02-07T21:23:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1802.02640v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1802.02640", "identifiers": {"arxiv": "1802.02640"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1802.02640"}], "name": "Minimizing Latency for Secure Coded Computing Using Secret Sharing via Staircase Codes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-02-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-02-07", "url": "https://arxiv.org/abs/1802.02640"}
{"aliases": [], "authors": ["Peng Zhang", "Yueming Liu", "Meikang Qiu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60cfaccc87a976654d30fc98aac27038a5f1c1c63a55b907f7c9bf3c257f79ca", "submitted_at": "2018-02-09T20:20:14Z", "updated_at": "2018-02-09T20:20:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1802.04766v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1802.04766", "identifiers": {"arxiv": "1802.04766"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1802.04766"}], "name": "SNC: A Cloud Service Platform for Symbolic-Numeric Computation using Just-In-Time Compilation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-02-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-02-09", "url": "https://arxiv.org/abs/1802.04766"}
{"aliases": [], "authors": ["David Goluskin", "Giovanni Fantuzzi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "907dc638cfa81c3a56af5795bf07b643b04aee7103a9d27c42d7f2ee84fd8bff", "submitted_at": "2018-02-22T18:52:32Z", "updated_at": "2018-09-06T18:23:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1802.08240v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1802.08240", "identifiers": {"arxiv": "1802.08240"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1802.08240"}], "name": "Bounds on mean energy in the Kuramoto-Sivashinsky equation computed using semidefinite programming", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-02-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-09-06", "url": "https://arxiv.org/abs/1802.08240"}
{"aliases": [], "authors": ["Evan Zheran Liu", "Kelvin Guu", "Panupong Pasupat", "Tianlin Shi", "Percy Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75d19fa97ce85a93bb9665a8ae06607bc0b90fde1c8376cdd60e99b55e90729c", "submitted_at": "2018-02-24T05:32:47Z", "updated_at": "2018-02-24T05:32:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1802.08802v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:1802.08802", "identifiers": {"arxiv": "1802.08802"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/1802.08802"}], "name": "Reinforcement Learning on Web Interfaces Using Workflow-Guided Exploration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2018-02-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses workflows extracted from demonstrations to constrain reinforcement-learning exploration for interactive web tasks.", "topics": ["reinforcement-learning", "workflow"], "updated_at": "2018-02-24", "url": "https://arxiv.org/abs/1802.08802"}
{"aliases": [], "authors": ["Kalev Alpernas", "Cormac Flanagan", "Sadjad Fouladi", "Leonid Ryzhyk", "Mooly Sagiv", "Thomas Schmitz", "Keith Winstein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "218feae88402c4aa28ef9f446a7cc752212c7d224baf4d9a284ff8ea3d5fbea9", "submitted_at": "2018-02-25T10:36:56Z", "updated_at": "2018-02-25T10:36:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1802.08984v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1802.08984", "identifiers": {"arxiv": "1802.08984"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1802.08984"}], "name": "Secure Serverless Computing Using Dynamic Information Flow Control", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-02-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-02-25", "url": "https://arxiv.org/abs/1802.08984"}
{"aliases": [], "authors": ["Mathias Soeken", "Thomas Häner", "Martin Roetteler"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68f2d0ad784d0bd4d8b91fb14c3a004b2e85ca0586882049fd894ab6812de98e", "submitted_at": "2018-03-02T19:41:14Z", "updated_at": "2018-03-02T19:41:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1803.01022v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1803.01022", "identifiers": {"arxiv": "1803.01022"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1803.01022"}], "name": "Programming Quantum Computers Using Design Automation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-03-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-03-02", "url": "https://arxiv.org/abs/1803.01022"}
{"aliases": [], "authors": ["Mordechai Guri", "Yosef Solewicz", "Andrey Daidakulov", "Yuval Elovici"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cb851ac90a57feee2726280d9304d750aaa353d1c905d0922080e0f7927a8aa5", "submitted_at": "2018-03-09T09:01:30Z", "updated_at": "2018-03-16T10:38:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1803.03422v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1803.03422", "identifiers": {"arxiv": "1803.03422"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1803.03422"}], "name": "MOSQUITO: Covert Ultrasonic Transmissions between Two Air-Gapped Computers using Speaker-to-Speaker Communication", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-03-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-03-16", "url": "https://arxiv.org/abs/1803.03422"}
{"aliases": [], "authors": ["Malik Hassanaly", "Venkat Raman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9d493eb3e359d5077262f00b2b9d95935ccadfaf7102d02867e23d9d9ddbceb", "submitted_at": "2018-05-01T17:08:34Z", "updated_at": "2018-05-01T17:08:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1805.00824v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1805.00824", "identifiers": {"arxiv": "1805.00824"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1805.00824"}], "name": "Numerical convergence of the Lyapunov spectrum computed using low Mach number solvers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-05-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-05-01", "url": "https://arxiv.org/abs/1805.00824"}
{"aliases": [], "authors": ["Yaoqing Yang", "Malhar Chaudhari", "Pulkit Grover", "Soummya Kar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2411a453a99497bf460fba9807f8a9886b9400253f015fc62256ce10ef2cf2a8", "submitted_at": "2018-05-15T22:08:05Z", "updated_at": "2018-05-15T22:08:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1805.06046v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1805.06046", "identifiers": {"arxiv": "1805.06046"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1805.06046"}], "name": "Coded Iterative Computing using Substitute Decoding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-05-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-05-15", "url": "https://arxiv.org/abs/1805.06046"}
{"aliases": [], "authors": ["Anirban Pathak"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9917ea2c7d786b29891f611416894793190a5c7bf6559a3c22698b7a967e9510", "submitted_at": "2018-05-16T12:45:31Z", "updated_at": "2018-05-16T12:45:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1805.06275v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1805.06275", "identifiers": {"arxiv": "1805.06275"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1805.06275"}], "name": "Experimental quantum mechanics in the class room: Testing basic ideas of quantum mechanics and quantum computing using IBM quantum computer", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-05-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-05-16", "url": "https://arxiv.org/abs/1805.06275"}
{"aliases": [], "authors": ["Jonny O'Dwyer", "Niall Murray", "Ronan Flynn"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3ffca92e9881bfba42bfa5cbf9de984a5a1aa5328913f076d55daac13f3b9701", "submitted_at": "2018-05-17T08:34:49Z", "updated_at": "2018-05-17T08:34:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1805.06652v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1805.06652", "identifiers": {"arxiv": "1805.06652"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1805.06652"}], "name": "Affective computing using speech and eye gaze: a review and bimodal system proposal for continuous affect prediction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-05-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-05-17", "url": "https://arxiv.org/abs/1805.06652"}
{"aliases": [], "authors": ["Carl Witt", "Marc Bux", "Wladislaw Gusew", "Ulf Leser"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e6c51fce6f2e7b309e64580436e5f789ab0f3fbbfc5d2cd7e40e9e0b59f685a", "submitted_at": "2018-05-30T09:24:08Z", "updated_at": "2018-05-30T09:24:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1805.11877v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1805.11877", "identifiers": {"arxiv": "1805.11877"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1805.11877"}], "name": "Predictive Performance Modeling for Distributed Computing using Black-Box Monitoring and Machine Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-05-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-05-30", "url": "https://arxiv.org/abs/1805.11877"}
{"aliases": [], "authors": ["Zhenghao Peng", "Xuyang Chen", "Chengwen Xu", "Naifeng Jing", "Xiaoyao Liang", "Cewu Lu", "Li Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c03a8f92490d4c7aa414a5f4e0286683f035f1befe9016a58fe4d76af18870f6", "submitted_at": "2018-07-27T06:59:46Z", "updated_at": "2018-12-18T08:57:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1807.10458v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1807.10458", "identifiers": {"arxiv": "1807.10458"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1807.10458"}], "name": "AXNet: ApproXimate computing using an end-to-end trainable neural network", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-07-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-12-18", "url": "https://arxiv.org/abs/1807.10458"}
{"aliases": [], "authors": ["Christopher J. Geoga", "Mihai Anitescu", "Michael L. Stein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3f397431751b9ece04b8040cc0e430e615d3e3659b74bcb33f37a5fa78f8e5d", "submitted_at": "2018-08-09T16:04:25Z", "updated_at": "2019-03-22T19:50:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1808.03215v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1808.03215", "identifiers": {"arxiv": "1808.03215"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1808.03215"}], "name": "Scalable Gaussian Process Computations Using Hierarchical Matrices", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-08-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-03-22", "url": "https://arxiv.org/abs/1808.03215"}
{"aliases": [], "authors": ["Andreea Buga", "Sorana Tania Nemes", "Atif Mashkoor"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6befbb4701dd818e9951d687e266834bcd75bbea46305874bbbc13f9f95c5592", "submitted_at": "2018-08-13T13:59:38Z", "updated_at": "2018-08-13T13:59:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1808.04222v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1808.04222", "identifiers": {"arxiv": "1808.04222"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1808.04222"}], "name": "Addressing Client Needs for Cloud Computing using Formal Foundations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-08-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-08-13", "url": "https://arxiv.org/abs/1808.04222"}
{"aliases": [], "authors": ["Panupong Pasupat", "Tian-Shun Jiang", "Evan Zheran Liu", "Kelvin Guu", "Percy Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f368116f70e987cde111835a05b5b36e0d345dabb4e8351610e2bbfb691b3440", "submitted_at": "2018-08-28T06:09:39Z", "updated_at": "2018-10-01T03:22:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1808.09132v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:1808.09132", "identifiers": {"arxiv": "1808.09132"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/1808.09132"}], "name": "Mapping Natural Language Commands to Web Elements", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2018-08-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Defines language grounding on webpages as choosing the UI element targeted by a natural-language command. A collection of more than 50,000 commands and three baseline models study functional references, relationships between elements, and visual position cues.", "topics": ["grounding", "data-generation", "evaluation"], "updated_at": "2018-10-01", "url": "https://arxiv.org/abs/1808.09132"}
{"aliases": [], "authors": ["Dennis Ehlers", "Markku Suomalainen", "Jens Lundell", "Ville Kyrki"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9979fd633906ae5d1e05a8df1d0f59b0f85ab155f4f1a4aa7b59d0646a46fac2", "submitted_at": "2018-09-13T09:49:24Z", "updated_at": "2019-03-25T13:06:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/1809.04860v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:1809.04860", "identifiers": {"arxiv": "1809.04860"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1809.04860"}], "name": "Imitating Human Search Strategies for Assembly", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-09-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-03-25", "url": "https://arxiv.org/abs/1809.04860"}
{"aliases": [], "authors": ["Alexandre Gélin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc926400155a41d1ffc4f5bcdea4e1ab5f3737cd469f8480bd033db796eded7c", "submitted_at": "2018-10-29T09:10:39Z", "updated_at": "2018-10-29T09:10:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1810.12010v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1810.12010", "identifiers": {"arxiv": "1810.12010"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1810.12010"}], "name": "Reducing the complexity for class group computations using small defining polynomials", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-10-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-10-29", "url": "https://arxiv.org/abs/1810.12010"}
{"aliases": [], "authors": ["David Diego", "Kristian Agasøster Haaga", "Bjarte Hannisdal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "22296e71f90e80313a34388538d7db35e8a7ba1a438afcf7357dbf303b7a8f0c", "submitted_at": "2018-11-05T13:40:47Z", "updated_at": "2019-04-02T16:25:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1811.01677v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1811.01677", "identifiers": {"arxiv": "1811.01677"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1811.01677"}], "name": "Transfer entropy computation using the Perron-Frobenius operator", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-11-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-04-02", "url": "https://arxiv.org/abs/1811.01677"}
{"aliases": [], "authors": ["Kleber Vieira", "Fernando Koch", "Joao Bosco Mangueira Sobral", "Carlos Becker Westphall", "Jorge Lopes de Souza Leao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "05f0414eb54439b39381d3f53c37aa2e24efbf8d2b2a10a9be05d6533850a284", "submitted_at": "2018-11-13T16:58:48Z", "updated_at": "2018-11-13T16:58:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1811.05407v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1811.05407", "identifiers": {"arxiv": "1811.05407"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1811.05407"}], "name": "Autonomic Intrusion Response in Distributed Computing using Big Data", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-11-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-11-13", "url": "https://arxiv.org/abs/1811.05407"}
{"aliases": [], "authors": ["Borja Ibarz", "Jan Leike", "Tobias Pohlen", "Geoffrey Irving", "Shane Legg", "Dario Amodei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d19839b8c60b94d4e6b12973bdf92cfd9aaa74d5140ecdc97b133793bcc15177", "submitted_at": "2018-11-15T18:33:43Z", "updated_at": "2018-11-15T18:33:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1811.06521v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:1811.06521", "identifiers": {"arxiv": "1811.06521"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1811.06521"}], "name": "Reward learning from human preferences and demonstrations in Atari", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-11-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-11-15", "url": "https://arxiv.org/abs/1811.06521"}
{"aliases": [], "authors": ["Jan Leike", "David Krueger", "Tom Everitt", "Miljan Martic", "Vishal Maini", "Shane Legg"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3cc6d0f663f0d3305f1bc64fe16551ab3b4fd378b30a221cf03785f4f933ca79", "submitted_at": "2018-11-19T18:48:04Z", "updated_at": "2018-11-19T18:48:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1811.07871v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:1811.07871", "identifiers": {"arxiv": "1811.07871"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1811.07871"}], "name": "Scalable agent alignment via reward modeling: a research direction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-11-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-11-19", "url": "https://arxiv.org/abs/1811.07871"}
{"aliases": [], "authors": ["Xinpeng Chen", "Lin Ma", "Jingyuan Chen", "Zequn Jie", "Wei Liu", "Jiebo Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb8925ccd2b4afa278f7730b0d0892b907317b3466d3da56dbc085a50eaf4e86", "submitted_at": "2018-12-09T04:30:11Z", "updated_at": "2018-12-09T04:30:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1812.03426v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:1812.03426", "identifiers": {"arxiv": "1812.03426"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1812.03426"}], "name": "Real-Time Referring Expression Comprehension by Single-Stage Grounding Network", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2018-12-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2018-12-09", "url": "https://arxiv.org/abs/1812.03426"}
{"aliases": [], "authors": ["Reza Akbarinia", "Bertrand Cloez"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b45590ec9c42f765897551719914a8bc6c62394d0ee22ddce4954719acba1e49", "submitted_at": "2019-01-17T10:05:37Z", "updated_at": "2019-01-17T10:05:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1901.05708v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1901.05708", "identifiers": {"arxiv": "1901.05708"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1901.05708"}], "name": "Efficient Matrix Profile Computation Using Different Distance Functions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-01-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-01-17", "url": "https://arxiv.org/abs/1901.05708"}
{"aliases": [], "authors": ["Mark J. Meyer", "Jeffrey S. Morris", "Regina Paxton Gazes", "Brent A. Coull"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c657c6d96dd85ba7793cbe38a6736f71e33257ca0f51cb4954643c1b2463b41", "submitted_at": "2019-01-23T16:17:20Z", "updated_at": "2021-03-18T15:18:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1901.07976v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1901.07976", "identifiers": {"arxiv": "1901.07976"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1901.07976"}], "name": "Ordinal Probit Functional Outcome Regression with Application to Computer-Use Behavior in Rhesus Monkeys", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-01-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-03-18", "url": "https://arxiv.org/abs/1901.07976"}
{"aliases": [], "authors": ["Kory W. Mathewson", "Pablo Samuel Castro", "Colin Cherry", "George Foster", "Marc G. Bellemare"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "59f37431469cceec9bd48d66f578b92c6502afc0d1515e371d6bd4f98056780b", "submitted_at": "2019-01-31T18:48:19Z", "updated_at": "2019-01-31T18:48:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1901.11528v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:1901.11528", "identifiers": {"arxiv": "1901.11528"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1901.11528"}], "name": "Shaping the Narrative Arc: An Information-Theoretic Approach to Collaborative Dialogue", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-01-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-01-31", "url": "https://arxiv.org/abs/1901.11528"}
{"aliases": [], "authors": ["Vincent T. Lee", "Samuel Archibald Elliot", "Armin Alaghi", "Luis Ceze"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4a0be3888d768c7145d2681d6c9beda5908f06f458ddc1bf4c1fff1ed8b32b8d", "submitted_at": "2019-02-15T19:23:30Z", "updated_at": "2019-02-26T21:08:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1902.05971v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1902.05971", "identifiers": {"arxiv": "1902.05971"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1902.05971"}], "name": "Synthesizing Number Generators for Stochastic Computing using Mixed Integer Programming", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-02-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-02-26", "url": "https://arxiv.org/abs/1902.05971"}
{"aliases": [], "authors": ["Venkata Pavan Kumar Miriyala", "Zhifeng Zhu", "Gengchiau Liang", "Xuanyao Fong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "66ea8c2a5dd45879558e1fa5e77825e37482e1322f18bd33e183ec8baf4de20c", "submitted_at": "2019-02-19T07:02:15Z", "updated_at": "2019-05-21T01:31:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1902.06925v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1902.06925", "identifiers": {"arxiv": "1902.06925"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1902.06925"}], "name": "Spin-wave mediated interactions for Majority Computation using Skyrmions and Spin-torque Nano-oscillators", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-02-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-05-21", "url": "https://arxiv.org/abs/1902.06925"}
{"aliases": [], "authors": ["Antoine Paris", "Hamed Mirghasemi", "Ivan Stupia", "Luc Vandendorpe"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8e71e656309f2fb7376a52b5f853a42f9001e501cd59528817f13e297ca7518", "submitted_at": "2019-03-06T10:33:38Z", "updated_at": "2019-03-06T10:33:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1903.02294v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1903.02294", "identifiers": {"arxiv": "1903.02294"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1903.02294"}], "name": "Energy-Efficient Edge-Facilitated Wireless Collaborative Computing using Map-Reduce", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-03-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-03-06", "url": "https://arxiv.org/abs/1903.02294"}
{"aliases": [], "authors": ["Ryan E. Scott", "Paul M. Alsing", "A. Matthew Smith", "Michael L. Fanto", "Christopher C. Tison", "James Schneeloch", "Edwin E. Hach,"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60030ee417f0a313d36cc6268699c1f2b65adb89ea7f6f797109e37c11800c54", "submitted_at": "2019-04-03T23:29:50Z", "updated_at": "2019-04-03T23:29:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1904.02268v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1904.02268", "identifiers": {"arxiv": "1904.02268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1904.02268"}], "name": "A scalable Controlled NOT gate for linear optical computing using microring resonators", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-04-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-04-03", "url": "https://arxiv.org/abs/1904.02268"}
{"aliases": [], "authors": ["Rahim K. Charania"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0b33f06099d0f7eb1bafe8091a034ee0a14849230d9b8c53cadf628824d870c", "submitted_at": "2019-04-06T02:43:30Z", "updated_at": "2019-04-06T02:43:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1904.03343v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1904.03343", "identifiers": {"arxiv": "1904.03343"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1904.03343"}], "name": "Exploring and Benchmarking High Performance & Scientific Computing using R R HPC Packages and Lower level compiled languages A Comparative Study", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-04-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-04-06", "url": "https://arxiv.org/abs/1904.03343"}
{"aliases": [], "authors": ["Jianping Zhu", "Rui Hou", "XiaoFeng Wang", "Wenhao Wang", "Jiangfeng Cao", "Lutan Zhao", "Fengkai Yuan", "Peinan Li", "Zhongpu Wang", "Boyan Zhao", "Lixin Zhang", "Dan Meng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f1c88da181a4409d49950dbeda6aac1e1fe78cdfafd310fb08ffde87ca4003d1", "submitted_at": "2019-04-09T16:56:21Z", "updated_at": "2019-04-10T07:19:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1904.04782v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1904.04782", "identifiers": {"arxiv": "1904.04782"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1904.04782"}], "name": "Enabling Privacy-Preserving, Compute- and Data-Intensive Computing using Heterogeneous Trusted Execution Environment", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-04-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-04-10", "url": "https://arxiv.org/abs/1904.04782"}
{"aliases": [], "authors": ["Akira Imakura", "Keiichi Morikuni", "Akitoshi Takayasu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50f726189fa6fd29ea23554e457a731fdc91b00d289dfe8c3e084dc1c86e718e", "submitted_at": "2019-04-12T15:29:35Z", "updated_at": "2019-09-25T02:51:35Z", "version": 3, "versioned_url": "https://arxiv.org/abs/1904.06277v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1904.06277", "identifiers": {"arxiv": "1904.06277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1904.06277"}], "name": "Verified partial eigenvalue computations using contour integrals for Hermitian generalized eigenproblems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-04-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-09-25", "url": "https://arxiv.org/abs/1904.06277"}
{"aliases": [], "authors": ["Erick Lavoie", "Laurie Hendren", "Fréderic Desprez", "Miguel Correia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb2cb198f15e0049b286283eaff49c6e69415ddd76afc3d9ab7bf80bc5c088b2", "submitted_at": "2019-04-25T15:20:28Z", "updated_at": "2019-04-25T15:20:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1904.11402v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1904.11402", "identifiers": {"arxiv": "1904.11402"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1904.11402"}], "name": "Genet: A Quickly Scalable Fat-Tree Overlay for Personal Volunteer Computing using WebRTC", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-04-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-04-25", "url": "https://arxiv.org/abs/1904.11402"}
{"aliases": [], "authors": ["Craig Gidney", "Austin G. Fowler"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "71b006acd113660b66f2e008288015f29a0990f8793bdac1273e7f52349e662d", "submitted_at": "2019-05-22T01:44:12Z", "updated_at": "2019-05-22T01:44:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1905.08916v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1905.08916", "identifiers": {"arxiv": "1905.08916"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1905.08916"}], "name": "Flexible layout of surface code computations using AutoCCZ states", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-05-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-05-22", "url": "https://arxiv.org/abs/1905.08916"}
{"aliases": [], "authors": ["Takahiro Maekawa", "Ayana Kawamura", "Takayuki Nakachi", "Hitoshi Kiya"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4333c5305d850e58c014b54de83bc2fbaabe2bce6bf2d9cf3320b13b6a9b11b8", "submitted_at": "2019-08-20T08:05:18Z", "updated_at": "2019-08-20T08:05:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1908.07915v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1908.07915", "identifiers": {"arxiv": "1908.07915"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1908.07915"}], "name": "Privacy-Preserving Support Vector Machine Computing Using Random Unitary Transformation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-08-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-08-20", "url": "https://arxiv.org/abs/1908.07915"}
{"aliases": [], "authors": ["Toby Jia-Jun Li", "Marissa Radensky", "Justin Jia", "Kirielle Singarajah", "Tom M. Mitchell", "Brad A. Myers"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4cef465f53abcbaad50c83f24e0ce54fb7f23390955e9f7e5592ec0b8287839e", "submitted_at": "2019-08-30T18:35:01Z", "updated_at": "2020-01-07T00:57:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1909.00031v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:46Z", "role": "primary", "sha256": "ee42fc28e088432e79837641c310dbd2d72f260a2a01d86f5f7f8b9c1b544a39", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/1909.00031v2"}], "id": "arxiv:1909.00031", "identifiers": {"arxiv": "1909.00031"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:46Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/1909.00031"}], "name": "Interactive Task and Concept Learning from Natural Language Instructions and GUI Demonstrations", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2019-08-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PUMICE, an agent that learns mobile-app procedures and conditional concepts through natural-language conversation and GUI demonstrations, recursively resolving ambiguous instructions with users.", "topics": ["human-in-the-loop", "mobile-navigation", "imitation-learning"], "updated_at": "2020-01-07", "url": "https://arxiv.org/abs/1909.00031"}
{"aliases": [], "authors": ["Alex Devonport", "Murat Arcak"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7dcf36ada74bcb7490e35df3038c3df5651b55556606a601d78c3f7f4bcbfcd5", "submitted_at": "2019-10-06T18:39:57Z", "updated_at": "2019-10-06T18:39:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1910.02500v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1910.02500", "identifiers": {"arxiv": "1910.02500"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1910.02500"}], "name": "Data-Driven Reachable Set Computation using Adaptive Gaussian Process Classification and Monte Carlo Methods", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-10-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-10-06", "url": "https://arxiv.org/abs/1910.02500"}
{"aliases": [], "authors": ["Ilan Tzitrin", "J. Eli Bourassa", "Nicolas C. Menicucci", "Krishna Kumar Sabapathy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f2489e3f61201b6f8c40ffbc0c97c7b3f2480d7af2a9a259617e6fe964b07c4", "submitted_at": "2019-10-08T20:32:27Z", "updated_at": "2020-03-16T14:18:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1910.03673v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1910.03673", "identifiers": {"arxiv": "1910.03673"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1910.03673"}], "name": "Progress towards practical qubit computation using approximate Gottesman-Kitaev-Preskill codes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-10-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-03-16", "url": "https://arxiv.org/abs/1910.03673"}
{"aliases": [], "authors": ["Nathan Berkovits", "Luis Alberto Ypanaqué"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9fd6c78bb6dbad4d5ced6ed92aeeec5c2530af73a18576543584a1511c23d261", "submitted_at": "2019-10-23T12:07:34Z", "updated_at": "2019-10-23T12:07:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1910.10511v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1910.10511", "identifiers": {"arxiv": "1910.10511"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1910.10511"}], "name": "Topological Amplitude Computations using the Pure Spinor Formalism", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-10-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-10-23", "url": "https://arxiv.org/abs/1910.10511"}
{"aliases": [], "authors": ["Dennis Núñez Fernández"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e331c5bcae6dad7380909a392cf6ddbd418307bfef5a572a2c1dfe56edbd5d2c", "submitted_at": "2019-10-18T08:09:23Z", "updated_at": "2019-10-18T08:09:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1910.11100v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:1910.11100", "identifiers": {"arxiv": "1910.11100"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1910.11100"}], "name": "Development of a hand pose recognition system on an embedded computer using CNNs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-10-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2019-10-18", "url": "https://arxiv.org/abs/1910.11100"}
{"aliases": [], "authors": ["Jordi Pont-Tuset", "Jasper Uijlings", "Soravit Changpinyo", "Radu Soricut", "Vittorio Ferrari"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e370bd79e2b80f3f18713989448c794a51ea56dfde61b4c1708aa3b14bf6cb6e", "submitted_at": "2019-12-06T13:21:16Z", "updated_at": "2020-07-20T17:18:38Z", "version": 4, "versioned_url": "https://arxiv.org/abs/1912.03098v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:1912.03098", "identifiers": {"arxiv": "1912.03098"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/1912.03098"}], "name": "Connecting Vision and Language with Localized Narratives", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2019-12-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-07-20", "url": "https://arxiv.org/abs/1912.03098"}
{"aliases": [], "authors": ["Kamran Alipour", "Jurgen P. Schulze", "Yi Yao", "Avi Ziskind", "Giedrius Burachas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f5b0538b0d8eb24bd3522138f92137a1eff8644e2363c5705e25018d601ba529", "submitted_at": "2020-03-01T07:54:01Z", "updated_at": "2020-03-01T07:54:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2003.00431v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2003.00431", "identifiers": {"arxiv": "2003.00431"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2003.00431"}], "name": "A Study on Multimodal and Interactive Explanations for Visual Question Answering", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-03-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-03-01", "url": "https://arxiv.org/abs/2003.00431"}
{"aliases": [], "authors": ["Marc W. Howard", "Michael E. Hasselmo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c9c5d76c22a3d01392194e7dbf9837cd183c2dfeac4766282ba0a3009f43f4fd", "submitted_at": "2020-03-25T22:40:49Z", "updated_at": "2020-03-25T22:40:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2003.11668v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2003.11668", "identifiers": {"arxiv": "2003.11668"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2003.11668"}], "name": "Cognitive computation using neural representations of time and space in the Laplace domain", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-03-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-03-25", "url": "https://arxiv.org/abs/2003.11668"}
{"aliases": [], "authors": ["Behzad Hasani", "Pooran Singh Negi", "Mohammad H. Mahoor"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "325421d070c8a503b0beb13c5c1d0be0d989335027874269e754ab88e940856f", "submitted_at": "2020-04-18T00:26:36Z", "updated_at": "2020-04-18T00:26:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2004.08495v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2004.08495", "identifiers": {"arxiv": "2004.08495"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2004.08495"}], "name": "BReG-NeXt: Facial Affect Computing Using Adaptive Residual Networks With Bounded Gradient", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-04-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-04-18", "url": "https://arxiv.org/abs/2004.08495"}
{"aliases": [], "authors": ["Arjun Majumdar", "Ayush Shrivastava", "Stefan Lee", "Peter Anderson", "Devi Parikh", "Dhruv Batra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8678175cb3078fbbaeaa37e65059b283649c92c750ba1ea3954824f62e4a8f2b", "submitted_at": "2020-04-30T17:22:40Z", "updated_at": "2020-05-01T17:16:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2004.14973v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2004.14973", "identifiers": {"arxiv": "2004.14973"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2004.14973"}], "name": "Improving Vision-and-Language Navigation with Image-Text Pairs from the Web", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-04-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-05-01", "url": "https://arxiv.org/abs/2004.14973"}
{"aliases": [], "authors": ["Menno A. Veerman", "Robert Pincus", "Robin Stoffer", "Caspar van Leeuwen", "Damian Podareanu", "Chiel C. van Heerwaarden"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57e29ef0da84b40450999995e17459c4b73cf0b1f9262650753aa5c524913f11", "submitted_at": "2020-05-05T15:00:58Z", "updated_at": "2020-08-16T08:59:30Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2005.02265v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2005.02265", "identifiers": {"arxiv": "2005.02265"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2005.02265"}], "name": "Predicting atmospheric optical properties for radiative transfer computations using neural networks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-05-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-08-16", "url": "https://arxiv.org/abs/2005.02265"}
{"aliases": [], "authors": ["Yang Li", "Jiacong He", "Xin Zhou", "Yuan Zhang", "Jason Baldridge"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3af7b45455ffbd616e8cdf5612b19275ecb7bf07b1065af9f5b724ae19e96fb4", "submitted_at": "2020-05-07T21:41:40Z", "updated_at": "2020-06-05T02:11:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2005.03776v2"}, "publication": {"authors": ["Yang Li", "Jiacong He", "Xin Zhou", "Yuan Zhang", "Jason Baldridge"], "date_basis": "Explicit publisher metadata: 2020/7; precision=month.", "date_parts": [2020, 7], "doi": "10.18653/v1/2020.acl-main.729", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2020.acl-main.729/"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "8198-8210", "pdf_url": "https://aclanthology.org/2020.acl-main.729.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Mapping Natural Language Instructions to Mobile UI Action Sequences", "type": "conference-paper", "url": "https://aclanthology.org/2020.acl-main.729/", "venue": "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "c196294e19378d08d71a7cfd41c2e368f5a6b49ee7c3601257537b9ae5c42add", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2020.acl-main.729/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2005.03776", "identifiers": {"arxiv": "2005.03776"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2005.03776"}, {"rel": "publication", "url": "https://aclanthology.org/2020.acl-main.729/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2020.acl-main.729.pdf"}], "name": "Mapping Natural Language Instructions to Mobile UI Action Sequences", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2020-05-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PIXELHELP for predicting mobile UI action sequences from instructions, together with action-phrase annotations and synthetic UI action descriptions for training. Separate Transformers extract instruction steps and ground object descriptions using interface content and screen position.", "topics": ["mobile-navigation", "grounding", "data-generation", "action-prediction"], "updated_at": "2020-06-05", "url": "https://arxiv.org/abs/2005.03776"}
{"aliases": [], "authors": ["Eugene Bagdasaryan", "Vitaly Shmatikov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4be54044f08b380ff5a2dda7a09a4a1e6382487e2d07ab10338034f3771a96f2", "submitted_at": "2020-05-08T02:15:53Z", "updated_at": "2021-02-19T04:45:28Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2005.03823v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2005.03823", "identifiers": {"arxiv": "2005.03823"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2005.03823"}], "name": "Blind Backdoors in Deep Learning Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-05-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-02-19", "url": "https://arxiv.org/abs/2005.03823"}
{"aliases": [], "authors": ["Shunsuke Horii"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8d29fadfc61192e151fcc1ae02d9643fdfe50c00a7a1bd049bd228ad2ff7d48", "submitted_at": "2020-05-13T02:11:54Z", "updated_at": "2020-05-13T02:11:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2005.06118v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2005.06118", "identifiers": {"arxiv": "2005.06118"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2005.06118"}], "name": "Improved Computation-Communication Trade-Off for Coded Distributed Computing using Linear Dependence of Intermediate Values", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-05-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-05-13", "url": "https://arxiv.org/abs/2005.06118"}
{"aliases": [], "authors": ["Ahmed Abid Moueddene", "Nader Khammassi", "Koen Bertels", "Carmen G. Almudever"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "66144b9c35b8bf0157e0b774a012c08a9146e33f8fddc95144e97ab8b97c8d59", "submitted_at": "2020-05-13T14:29:18Z", "updated_at": "2020-05-13T14:29:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2005.06337v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2005.06337", "identifiers": {"arxiv": "2005.06337"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2005.06337"}], "name": "Realistic simulation of quantum computation using unitary and measurement channels", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-05-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-05-13", "url": "https://arxiv.org/abs/2005.06337"}
{"aliases": [], "authors": ["Corey Lynch", "Pierre Sermanet"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d0ddee8c5f6aa8c915c6cca3f8dcc0a0d35e1e4d1799758610fdd8c3540c017", "submitted_at": "2020-05-15T17:08:50Z", "updated_at": "2021-07-07T23:43:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2005.07648v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T08:01:28Z", "role": "primary", "sha256": "822cec31fa0db7ffa390fda806d88c2f9c0fe4d328742e7008b662916e92373a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2005.07648%2C2206.04615%2C2311.12022%2C2311.16922%2C2502.13130%2C2506.21506%2C2511.11332&max_results=7"}], "id": "arxiv:2005.07648", "identifiers": {"arxiv": "2005.07648"}, "kind": "paper", "last_verified_at": "2026-10-18T08:01:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2005.07648"}], "name": "Language Conditioned Imitation Learning over Unstructured Data", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-05-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-07-07", "url": "https://arxiv.org/abs/2005.07648"}
{"aliases": [], "authors": ["Aishan Liu", "Tairan Huang", "Xianglong Liu", "Yitao Xu", "Yuqing Ma", "Xinyun Chen", "Stephen J. Maybank", "Dacheng Tao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "31a0b56d0efddd546d3a9da97903f5152cf969994a575ef4b2b311d78798889f", "submitted_at": "2020-05-19T01:38:47Z", "updated_at": "2020-11-17T14:13:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2005.09161v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2005.09161", "identifiers": {"arxiv": "2005.09161"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2005.09161"}], "name": "Spatiotemporal Attacks for Embodied Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-05-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-11-17", "url": "https://arxiv.org/abs/2005.09161"}
{"aliases": [], "authors": ["Nicholas Chancellor", "Robert Cumming", "Tim Thomas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a3d39bf7f85f18c442ebde594ba730dec4f2141d9ff35a25189dc6d4449953e", "submitted_at": "2020-06-10T14:23:37Z", "updated_at": "2020-06-10T14:23:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2006.05846v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2006.05846", "identifiers": {"arxiv": "2006.05846"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2006.05846"}], "name": "Toward a standardized methodology for constructing quantum computing use cases", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-06-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-06-10", "url": "https://arxiv.org/abs/2006.05846"}
{"aliases": [], "authors": ["Yuichi Sano"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "28e09af6f62d1a22c9a8e86b3b9c0398c75ebb4e733fbf58e1f2c8b20094c52c", "submitted_at": "2020-06-11T08:37:45Z", "updated_at": "2020-09-10T09:59:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2006.06255v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2006.06255", "identifiers": {"arxiv": "2006.06255"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2006.06255"}], "name": "Blind Quantum Computation Using a Circuit-Based Quantum Computer", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-06-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-09-10", "url": "https://arxiv.org/abs/2006.06255"}
{"aliases": [], "authors": ["Dan Hendrycks", "Steven Basart", "Norman Mu", "Saurav Kadavath", "Frank Wang", "Evan Dorundo", "Rahul Desai", "Tyler Zhu", "Samyak Parajuli", "Mike Guo", "Dawn Song", "Jacob Steinhardt", "Justin Gilmer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ed24b67a5b384a432eabad868fd67ce6883cc3a126a7078ad24b51d2b9fa8b25", "submitted_at": "2020-06-29T17:59:10Z", "updated_at": "2021-07-24T04:28:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2006.16241v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2006.16241", "identifiers": {"arxiv": "2006.16241"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2006.16241"}], "name": "The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-06-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-07-24", "url": "https://arxiv.org/abs/2006.16241"}
{"aliases": [], "authors": ["Baber Khalid", "Malihe Alikhani", "Michael Fellner", "Brian McMahan", "Matthew Stone"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f362a6a667671d472fce6909dd3396403e7acc5809ef9bbc8005fc864daf8621", "submitted_at": "2020-07-08T20:53:14Z", "updated_at": "2020-07-08T20:53:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2007.04428v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2007.04428", "identifiers": {"arxiv": "2007.04428"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2007.04428"}], "name": "Discourse Coherence, Reference Grounding and Goal Oriented Dialogue", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-07-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-07-08", "url": "https://arxiv.org/abs/2007.04428"}
{"aliases": [], "authors": ["Yiming Li", "Yong Jiang", "Zhifeng Li", "Shu-Tao Xia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cd32a8e864d1a3c92049d0f70e8e3f0e0f4be5b31f234d739c88e6732ff270bd", "submitted_at": "2020-07-17T04:09:20Z", "updated_at": "2022-02-16T06:39:39Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2007.08745v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2007.08745", "identifiers": {"arxiv": "2007.08745"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2007.08745"}], "name": "Backdoor Learning: A Survey", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-07-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-02-16", "url": "https://arxiv.org/abs/2007.08745"}
{"aliases": [], "authors": ["Jieshan Chen", "Mulong Xie", "Zhenchang Xing", "Chunyang Chen", "Xiwei Xu", "Liming Zhu", "Guoqiang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6352985532945fd562ce0382aef0e977ce1ec5909b6872e5fe048167e9732a0", "submitted_at": "2020-08-12T06:36:33Z", "updated_at": "2020-09-07T12:57:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2008.05132v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2008.05132", "identifiers": {"arxiv": "2008.05132"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2008.05132"}], "name": "Object Detection for Graphical User Interface: Old Fashioned or Deep Learning or a Combination?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2020-08-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares seven traditional and learned GUI element detectors on more than 50,000 interface images. The analysis motivates a coarse-to-fine non-text detector combined with a learned text detector to address GUI-specific localization requirements.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2020-09-07", "url": "https://arxiv.org/abs/2008.05132"}
{"aliases": [], "authors": ["Donald R. Short", "Jerome A. Orosz", "Gur Windmiller", "William F. Welsh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7d3b62598c344dfadd2ca129096a8aca338649bbed52fddb2842d98d4012084", "submitted_at": "2020-08-21T23:19:22Z", "updated_at": "2020-10-27T22:52:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2008.09711v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2008.09711", "identifiers": {"arxiv": "2008.09711"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2008.09711"}], "name": "Fast Transit Computation Using Tabulated Stellar Intensities", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-08-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-10-27", "url": "https://arxiv.org/abs/2008.09711"}
{"aliases": [], "authors": ["Franz Georg Fuchs", "Herman Øie Kolden", "Niels Henrik Aase", "Giorgio Sartor"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ca83b0b7ce4d0f1c6b4cd65dfa7853f4f33c0e3f6046e2edc0cf088147354e75", "submitted_at": "2020-09-02T14:15:55Z", "updated_at": "2020-11-09T21:35:28Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2009.01095v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2009.01095", "identifiers": {"arxiv": "2009.01095"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2009.01095"}], "name": "Efficient encoding of the weighted MAX k-CUT on a quantum computer using QAOA", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-09-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-11-09", "url": "https://arxiv.org/abs/2009.01095"}
{"aliases": [], "authors": ["Damir Vrabac", "Akshay Smit", "Rebecca Rojansky", "Yasodha Natkunam", "Ranjana H. Advani", "Andrew Y. Ng", "Sebastian Fernandez-Pol", "Pranav Rajpurkar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7df8b548b51aca03479978db0808a1a301f308626f3acff6cad8cc80f7c51474", "submitted_at": "2020-09-17T07:43:42Z", "updated_at": "2020-09-24T11:02:28Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2009.08123v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2009.08123", "identifiers": {"arxiv": "2009.08123"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2009.08123"}], "name": "DLBCL-Morph: Morphological features computed using deep learning for an annotated digital DLBCL image set", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-09-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-09-24", "url": "https://arxiv.org/abs/2009.08123"}
{"aliases": [], "authors": ["Yuning Du", "Chenxia Li", "Ruoyu Guo", "Xiaoting Yin", "Weiwei Liu", "Jun Zhou", "Yifan Bai", "Zilin Yu", "Yehua Yang", "Qingqing Dang", "Haoshuang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "222c26f82629db9e7298c21dd4a0a063e8af2a81b985220be69db99baba196fc", "submitted_at": "2020-09-21T14:57:18Z", "updated_at": "2020-10-15T14:21:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2009.09941v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T20:10:43Z", "role": "primary", "sha256": "5ef9f505666be9f0288ac54624d94903315baae1f474f19a1eb5b1a2e157694e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2009.09941&max_results=1"}], "id": "arxiv:2009.09941", "identifiers": {"arxiv": "2009.09941"}, "kind": "paper", "last_verified_at": "2026-10-21T20:10:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2009.09941"}], "name": "PP-OCR: A Practical Ultra Lightweight OCR System", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-09-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-10-15", "url": "https://arxiv.org/abs/2009.09941"}
{"aliases": [], "authors": ["Hung Le", "Svetha Venkatesh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "840c735abb180e2226220e8ce09c177b50492666647e071a851b0e6e91ff2f9c", "submitted_at": "2020-09-24T01:39:16Z", "updated_at": "2020-09-24T01:39:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2009.11443v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2009.11443", "identifiers": {"arxiv": "2009.11443"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2009.11443"}], "name": "Neurocoder: Learning General-Purpose Computation Using Stored Neural Programs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-09-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-09-24", "url": "https://arxiv.org/abs/2009.11443"}
{"aliases": [], "authors": ["Ismaeel Al Ridhawi", "Moayad Aloqaily", "Yaser Jararweh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "48e73cd9d3c058c01fca0ba5966816ed3ecab3a6a1555ad7ef3b4a82bbe5c1a2", "submitted_at": "2020-09-24T18:48:22Z", "updated_at": "2020-09-24T18:48:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2009.11901v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2009.11901", "identifiers": {"arxiv": "2009.11901"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2009.11901"}], "name": "An Incentive-Based Mechanism for Volunteer Computing using Blockchain", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-09-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-09-24", "url": "https://arxiv.org/abs/2009.11901"}
{"aliases": [], "authors": ["Manuel Grimm", "Adrian Beckert", "Gabriel Aeppli", "Markus Müller"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4422c6436dcdaae7894fd339a6c8d8586363917e7aec81ee52b5046363f37fc7", "submitted_at": "2020-09-29T16:20:00Z", "updated_at": "2020-09-29T16:20:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2009.14126v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2009.14126", "identifiers": {"arxiv": "2009.14126"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2009.14126"}], "name": "Universal quantum computing using electro-nuclear wavefunctions of rare-earth ions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-09-29", "url": "https://arxiv.org/abs/2009.14126"}
{"aliases": [], "authors": ["Yang Li", "Gang Li", "Luheng He", "Jingjie Zheng", "Hong Li", "Zhiwei Guan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68af0a0a4b605d0c968e8913232fde16e3eadf6361099e8fc3d015dcdb26302b", "submitted_at": "2020-10-08T22:56:03Z", "updated_at": "2020-10-08T22:56:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2010.04295v1"}, "publication": {"authors": ["Yang Li", "Gang Li", "Luheng He", "Jingjie Zheng", "Hong Li", "Zhiwei Guan"], "date_basis": "Explicit publisher metadata: 2020/11; precision=month.", "date_parts": [2020, 11], "doi": "10.18653/v1/2020.emnlp-main.443", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2020.emnlp-main.443/"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The retained arXiv abstract reports 162,859 phrases; the ACL abstract reports 162,860. This publication proposal does not overwrite the arXiv annotation or assert a unified dataset count. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "5495-5510", "pdf_url": "https://aclanthology.org/2020.emnlp-main.443.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Widget Captioning: Generating Natural Language Description for Mobile User Interface Elements", "type": "conference-paper", "url": "https://aclanthology.org/2020.emnlp-main.443/", "venue": "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "750e60b4d28e5b97a46595ea2280b520bf9edcc445b9931bad225e0028935c5e", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2020.emnlp-main.443/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2010.04295", "identifiers": {"arxiv": "2010.04295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2010.04295"}, {"rel": "publication", "url": "https://aclanthology.org/2020.emnlp-main.443/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2020.emnlp-main.443.pdf"}], "name": "Widget Captioning: Generating Natural Language Description for Mobile User Interface Elements", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2020-10-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces crowdsourced descriptions of mobile UI elements and studies multimodal models that generate useful widget captions.", "topics": ["screen-understanding", "grounding"], "updated_at": "2020-10-08", "url": "https://arxiv.org/abs/2010.04295"}
{"aliases": [], "authors": ["Francesco Croce", "Maksym Andriushchenko", "Vikash Sehwag", "Edoardo Debenedetti", "Nicolas Flammarion", "Mung Chiang", "Prateek Mittal", "Matthias Hein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8122c6c091cf7b43892ab46c379f3845303981b8836249a3fe427f573a71de76", "submitted_at": "2020-10-19T17:06:18Z", "updated_at": "2021-10-31T20:03:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2010.09670v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2010.09670", "identifiers": {"arxiv": "2010.09670"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2010.09670"}], "name": "RobustBench: a standardized adversarial robustness benchmark", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-10-31", "url": "https://arxiv.org/abs/2010.09670"}
{"aliases": [], "authors": ["Simon Stepputtis", "Joseph Campbell", "Mariano Phielipp", "Stefan Lee", "Chitta Baral", "Heni Ben Amor"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb83f67b55d48c5bfc48920107458e15bac56c0e7fe523b4e8f2c9a10b199aa5", "submitted_at": "2020-10-22T21:49:08Z", "updated_at": "2020-10-22T21:49:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2010.12083v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2010.12083", "identifiers": {"arxiv": "2010.12083"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2010.12083"}], "name": "Language-Conditioned Imitation Learning for Robot Manipulation Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-10-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-10-22", "url": "https://arxiv.org/abs/2010.12083"}
{"aliases": [], "authors": ["S. Omkar", "Y. S. Teo", "Seung-Woo Lee", "H. Jeong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "473acbcf33d4e2b58c92e42a129063daef5bbced4670a8048361565dd597cf3b", "submitted_at": "2020-11-09T06:05:44Z", "updated_at": "2020-11-09T06:05:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2011.04209v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2011.04209", "identifiers": {"arxiv": "2011.04209"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2011.04209"}], "name": "Highly photon loss tolerant quantum computing using hybrid qubits", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-11-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-11-09", "url": "https://arxiv.org/abs/2011.04209"}
{"aliases": [], "authors": ["Runhua Xu", "James Joshi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c526d1a627dbe5a03c3a82067133368ea68788b1f5480fc08ac783b8ccbc613b", "submitted_at": "2020-11-12T04:28:04Z", "updated_at": "2020-12-07T17:39:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2011.06191v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2011.06191", "identifiers": {"arxiv": "2011.06191"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2011.06191"}], "name": "Revisiting Secure Computation Using Functional Encryption: Opportunities and Research Directions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-11-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-12-07", "url": "https://arxiv.org/abs/2011.06191"}
{"aliases": [], "authors": ["M. F. Gonzalez-Zalba", "S. de Franceschi", "E. Charbon", "T. Meunier", "M. Vinet", "A. S. Dzurak"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "94bb1678ea49ca5f887a101c49271a1566f8f58a5596e8fa052ea3e6d982c649", "submitted_at": "2020-11-23T21:59:39Z", "updated_at": "2023-04-08T10:59:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2011.11753v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2011.11753", "identifiers": {"arxiv": "2011.11753"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2011.11753"}], "name": "Scaling silicon-based quantum computing using CMOS technology: State-of-the-art, Challenges and Perspectives", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-11-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-04-08", "url": "https://arxiv.org/abs/2011.11753"}
{"aliases": [], "authors": ["Christopher Chamberland", "Kyungjoo Noh", "Patricio Arrangoiz-Arriola", "Earl T. Campbell", "Connor T. Hann", "Joseph Iverson", "Harald Putterman", "Thomas C. Bohdanowicz", "Steven T. Flammia", "Andrew Keller", "Gil Refael", "John Preskill", "Liang Jiang", "Amir H. Safavi-Naeini", "Oskar Painter", "Fernando G. S. L. Brandão"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9579123a555581245446558157e6596405e1c40137ad5c7a7ab33b2a06c98793", "submitted_at": "2020-12-07T23:22:40Z", "updated_at": "2022-01-27T08:21:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2012.04108v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2012.04108", "identifiers": {"arxiv": "2012.04108"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2012.04108"}], "name": "Building a fault-tolerant quantum computer using concatenated cat codes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-12-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-01-27", "url": "https://arxiv.org/abs/2012.04108"}
{"aliases": [], "authors": ["Andrew Simon Johnson", "William Winlow"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "42104912771a1c5ec1487b8995367fa4ea5619c8c1f11d0dcb81450f467b8188", "submitted_at": "2020-12-04T08:00:23Z", "updated_at": "2020-12-04T08:00:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2012.06537v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2012.06537", "identifiers": {"arxiv": "2012.06537"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2012.06537"}], "name": "Does the brain function as a quantum phase computer using phase ternary computation?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2020-12-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2020-12-04", "url": "https://arxiv.org/abs/2012.06537"}
{"aliases": [], "authors": ["Zecheng He", "Srinivas Sunkara", "Xiaoxue Zang", "Ying Xu", "Lijuan Liu", "Nevan Wichers", "Gabriel Schubiner", "Ruby Lee", "Jindong Chen", "Blaise Agüera y Arcas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51fdddb9a3d8a6090875719a51b22008b1f43bc38c6cfe1713a49a52f7e8434c", "submitted_at": "2020-12-22T20:49:52Z", "updated_at": "2021-01-25T20:37:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2012.12350v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:47Z", "role": "primary", "sha256": "139375513dafe5d7cfd0f5843b2a3465294e2b54051cf1c964a0d0751166a469", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2012.12350v2"}], "id": "arxiv:2012.12350", "identifiers": {"arxiv": "2012.12350"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2012.12350"}], "name": "ActionBert: Leveraging User Actions for Semantic Understanding of User Interfaces", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2020-12-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Pretrains ActionBert on user interaction traces to combine visual, linguistic, and UI-structure features. It evaluates the resulting UI representations on tasks including icon classification and description-based component retrieval.", "topics": ["screen-understanding", "grounding", "action-prediction"], "updated_at": "2021-01-25", "url": "https://arxiv.org/abs/2012.12350"}
{"aliases": [], "authors": ["Touheed Anwar Atif", "S. Sandeep Pradhan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0b4c64cd64a7d878b266b3be28cdb4336766d884f3d86b4ee60fad65e43e7e04", "submitted_at": "2021-01-07T04:22:15Z", "updated_at": "2021-10-14T20:06:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2101.02360v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2101.02360", "identifiers": {"arxiv": "2101.02360"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2101.02360"}], "name": "Distributed Quantum Faithful Simulation and Function Computation Using Algebraic Structured Measurements", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-01-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-10-14", "url": "https://arxiv.org/abs/2101.02360"}
{"aliases": [], "authors": ["Xiaoyi Zhang", "Lilian de Greef", "Amanda Swearngin", "Samuel White", "Kyle Murray", "Lisa Yu", "Qi Shan", "Jeffrey Nichols", "Jason Wu", "Chris Fleizach", "Aaron Everitt", "Jeffrey P. Bigham"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6a5305254ea405b6297c0a3eee43d58fc028abaa1a538d03b175b39b9daccca", "submitted_at": "2021-01-13T05:56:15Z", "updated_at": "2021-01-13T05:56:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2101.04893v1"}}, "environments": ["ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2101.04893", "identifiers": {"arxiv": "2101.04893"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2101.04893"}], "name": "Screen Recognition: Creating Accessibility Metadata for Mobile Applications from Pixels", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2021-01-13", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Infers UI elements and accessibility metadata from mobile screenshots to support screen-reader interaction with iOS applications.", "topics": ["screen-understanding"], "updated_at": "2021-01-13", "url": "https://arxiv.org/abs/2101.04893"}
{"aliases": [], "authors": ["Xingyu Chen", "Zihan Zhao", "Lu Chen", "Danyang Zhang", "Jiabao Ji", "Ao Luo", "Yuxuan Xiong", "Kai Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5fdf4bc31edd633fd0858835db0afedeef5dbc2694c4511903976c338bd19b1b", "submitted_at": "2021-01-23T09:43:44Z", "updated_at": "2021-11-08T08:31:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2101.09465v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2101.09465", "identifiers": {"arxiv": "2101.09465"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2101.09465"}], "name": "WebSRC: A Dataset for Web-Based Structural Reading Comprehension", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2021-01-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces question answering over webpage text and structure, with aligned HTML, screenshots, and metadata for structural reading comprehension.", "topics": ["screen-understanding", "evaluation"], "updated_at": "2021-11-08", "url": "https://arxiv.org/abs/2101.09465"}
{"aliases": [], "authors": ["Tirthak Patel", "Devesh Tiwari"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "613756875bf96eff636d013d6bba333ff812229fa6e529d565c4d9e815d70719", "submitted_at": "2021-02-01T20:43:53Z", "updated_at": "2021-02-01T20:43:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2102.01153v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2102.01153", "identifiers": {"arxiv": "2102.01153"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2102.01153"}], "name": "DisQ: A Novel Quantum Output State Classification Method on IBM Quantum Computers using OpenPulse", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-02-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-02-01", "url": "https://arxiv.org/abs/2102.01153"}
{"aliases": [], "authors": ["Christian W. Bauer", "Marat Freytsis", "Benjamin Nachman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0bf7a0cec466fe5584b67e2200c450c09437a5ad4e5a832eb9c0be98a826272a", "submitted_at": "2021-02-09T19:00:00Z", "updated_at": "2021-02-09T19:00:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2102.05044v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2102.05044", "identifiers": {"arxiv": "2102.05044"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2102.05044"}], "name": "Simulating collider physics on quantum computers using effective field theories", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-02-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-02-09", "url": "https://arxiv.org/abs/2102.05044"}
{"aliases": [], "authors": ["Sabrina Kirrane"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "93e9974fafcc8d13c242c4787c1147b399ec345a965c3f7d6bee3760e6298527", "submitted_at": "2021-02-12T16:32:02Z", "updated_at": "2021-09-24T14:16:41Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2102.06607v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2102.06607", "identifiers": {"arxiv": "2102.06607"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2102.06607"}], "name": "Intelligent Software Web Agents: A Gap Analysis", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-02-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-09-24", "url": "https://arxiv.org/abs/2102.06607"}
{"aliases": [], "authors": ["Yukio Kawashima", "Erika Lloyd", "Marc P. Coons", "Yunseong Nam", "Shunji Matsuura", "Alejandro J. Garza", "Sonika Johri", "Lee Huntington", "Valentin Senicourt", "Andrii O. Maksymov", "Jason H. V. Nguyen", "Jungsang Kim", "Nima Alidoust", "Arman Zaribafiyan", "Takeshi Yamazaki"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0b07efd80e70afab32cca17945b1707d29ef6cb256f0b9095f1798d332af68fc", "submitted_at": "2021-02-14T01:47:52Z", "updated_at": "2021-09-28T05:39:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2102.07045v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2102.07045", "identifiers": {"arxiv": "2102.07045"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2102.07045"}], "name": "Optimizing Electronic Structure Simulations on a Trapped-ion Quantum Computer using Problem Decomposition", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-02-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-09-28", "url": "https://arxiv.org/abs/2102.07045"}
{"aliases": [], "authors": ["Prajnesh Kumar", "Mingwei Jin", "Ting Bu", "Santosh Kumar", "Yu-Ping Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "953df741630f66a845d7d7b83c76c51292714c93fe56fb05686fd638dc5d24c4", "submitted_at": "2021-02-11T23:59:14Z", "updated_at": "2021-02-11T23:59:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2102.09049v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2102.09049", "identifiers": {"arxiv": "2102.09049"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2102.09049"}], "name": "Efficient Reservoir Computing using Field Programmable Gate Array and Electro-optic Modulation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-02-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-02-11", "url": "https://arxiv.org/abs/2102.09049"}
{"aliases": [], "authors": ["Johannes Pekkilä", "Miikka S. Väisälä", "Maarit J. Käpylä", "Matthias Rheinhardt", "Oskar Lappi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3cfc9f29eb40b2694d19e952d999e0a923c802648b82faf8d70e9e72ec3858d8", "submitted_at": "2021-03-02T09:44:42Z", "updated_at": "2022-05-10T11:21:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2103.01597v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2103.01597", "identifiers": {"arxiv": "2103.01597"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2103.01597"}], "name": "Scalable communication for high-order stencil computations using CUDA-aware MPI", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-03-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-05-10", "url": "https://arxiv.org/abs/2103.01597"}
{"aliases": [], "authors": ["Laha Ale", "Ning Zhang", "Xiaojie Fang", "Xianfu Chen", "Shaohua Wu", "Longzhuang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e85ab7fbeb46a1cc1a31da380858975a8b4581c1eabcfc823a3a2a970353d467", "submitted_at": "2021-03-13T23:10:40Z", "updated_at": "2021-03-13T23:10:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2103.07811v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2103.07811", "identifiers": {"arxiv": "2103.07811"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2103.07811"}], "name": "Delay-aware and Energy-Efficient Computation Offloading in Mobile Edge Computing Using Deep Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-03-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-03-13", "url": "https://arxiv.org/abs/2103.07811"}
{"aliases": [], "authors": ["Nancy Xu", "Sam Masling", "Michael Du", "Giovanni Campagna", "Larry Heck", "James Landay", "Monica S Lam"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e00b8f1b34097455ece2cf97efabbb64f398e3fffdec13394f731096672dc654", "submitted_at": "2021-03-30T04:02:34Z", "updated_at": "2021-04-04T23:51:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2103.16057v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2103.16057", "identifiers": {"arxiv": "2103.16057"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2103.16057"}], "name": "Grounding Open-Domain Instructions to Automate Web Support Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2021-03-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces RUSS, which translates support instructions into a domain-specific language and grounds them in actionable web elements.", "topics": ["grounding", "web-navigation"], "updated_at": "2021-04-04", "url": "https://arxiv.org/abs/2103.16057"}
{"aliases": [], "authors": ["Abdelkrim Alahyane", "Mohamed El Fakir", "Saad Benjelloun", "Ikram Chairi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ec63237855db994402517fca0f1616a3db736d54f18b5fe62485bb5b992b0afb", "submitted_at": "2021-04-16T17:26:46Z", "updated_at": "2021-04-16T17:26:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2104.08244v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2104.08244", "identifiers": {"arxiv": "2104.08244"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2104.08244"}], "name": "Open data for Moroccan license plates for OCR applications : data collection, labeling, and model construction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-04-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-04-16", "url": "https://arxiv.org/abs/2104.08244"}
{"aliases": [], "authors": ["Andrea Burns", "Deniz Arsan", "Sanjna Agrawal", "Ranjitha Kumar", "Kate Saenko", "Bryan A. Plummer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da0efe4e827a9d5ac3a662b6d4a792bf6573eb596c61b72c3227cd556ff6afb7", "submitted_at": "2021-04-17T14:48:02Z", "updated_at": "2021-04-17T14:48:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2104.08560v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2104.08560", "identifiers": {"arxiv": "2104.08560"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2104.08560"}], "name": "Mobile App Tasks with Iterative Feedback (MoTIF): Addressing Task Feasibility in Interactive Visual Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-04-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-04-17", "url": "https://arxiv.org/abs/2104.08560"}
{"aliases": [], "authors": ["Shlomi Dolev", "Arseni Kalma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf9873a3d61a09e4a5210d6ae26f80d24784b2e3ed6a476074fd74430fb464ca", "submitted_at": "2021-05-04T21:03:06Z", "updated_at": "2021-05-04T21:03:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2105.01762v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2105.01762", "identifiers": {"arxiv": "2105.01762"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2105.01762"}], "name": "Verifiable Computing Using Computation Fingerprints Within FHE", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-05-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-05-04", "url": "https://arxiv.org/abs/2105.01762"}
{"aliases": [], "authors": ["Zafar Ahmad", "Rezaul Chowdhury", "Rathish Das", "Pramod Ganapathi", "Aaron Gregory", "Yimin Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77deaa89199e6a22e888b78c34373256c3f050e758ddba1d803bae5f53f1baea", "submitted_at": "2021-05-14T07:24:36Z", "updated_at": "2021-05-14T07:24:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2105.06676v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2105.06676", "identifiers": {"arxiv": "2105.06676"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2105.06676"}], "name": "Fast Stencil Computations using Fast Fourier Transforms", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-05-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-05-14", "url": "https://arxiv.org/abs/2105.06676"}
{"aliases": [], "authors": ["Jingwen Fu", "Xiaoyi Zhang", "Yuwang Wang", "Wenjun Zeng", "Sam Yang", "Grayson Hilliard"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8788e419d1a901f8ad9dd6089a12fe129baee764b1c285bbeb4ea60d7046afa2", "submitted_at": "2021-05-25T13:45:54Z", "updated_at": "2022-04-14T11:00:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2105.11941v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:47Z", "role": "primary", "sha256": "41c4da71e457ea23725a8070b0da64c234f03613841b95edea3a5f3c95dd8c68", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2105.11941v2"}], "id": "arxiv:2105.11941", "identifiers": {"arxiv": "2105.11941"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2105.11941"}], "name": "Understanding Mobile GUI: from Pixel-Words to Screen-Sentences", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2021-05-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Models mobile screens as atomic text and graphic components whose relationships are learned by a Screen Transformer. The work introduces RICO-PW annotations and evaluates metadata-free screen understanding on relation, clickability, retrieval, and app-classification tasks.", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2022-04-14", "url": "https://arxiv.org/abs/2105.11941"}
{"aliases": [], "authors": ["Daniel Toyama", "Philippe Hamel", "Anita Gergely", "Gheorghe Comanici", "Amelia Glaese", "Zafarali Ahmed", "Tyler Jackson", "Shibl Mourad", "Doina Precup"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f0caf694223bc2569cdd01662cadf54cbc96cad370866a231c15437c5d5d0ba", "submitted_at": "2021-05-27T15:20:14Z", "updated_at": "2021-05-27T15:20:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2105.13231v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2105.13231", "identifiers": {"arxiv": "2105.13231"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2105.13231"}], "name": "AndroidEnv: A Reinforcement Learning Platform for Android", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2021-05-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents an Android environment for training and evaluating reinforcement-learning agents through touchscreen interaction.", "topics": ["reinforcement-learning", "sandbox"], "updated_at": "2021-05-27", "url": "https://arxiv.org/abs/2105.13231"}
{"aliases": [], "authors": ["M. Mazzanti", "R. X. Schüssler", "J. D. Arias Espinoza", "Z. Wu", "R. Gerritsma", "A. Safavi-Naini"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6ec85c314761779632c44e05601644cf1e150618c0aba2cd4e2743c41e1892f4", "submitted_at": "2021-06-14T15:16:16Z", "updated_at": "2021-06-14T15:16:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2106.07486v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2106.07486", "identifiers": {"arxiv": "2106.07486"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2106.07486"}], "name": "Trapped Ion Quantum Computing using Optical Tweezers and Electric Fields", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-06-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-06-14", "url": "https://arxiv.org/abs/2106.07486"}
{"aliases": [], "authors": ["Chongkai Gao", "Haichuan Gao", "Shangqi Guo", "Tianren Zhang", "Feng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4609d1436d1ac58afbcf9337ad66c1d4b7f8f9569d8b8dfe645598897a00efa", "submitted_at": "2021-06-17T12:15:57Z", "updated_at": "2021-07-02T07:17:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2106.09422v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2106.09422", "identifiers": {"arxiv": "2106.09422"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2106.09422"}], "name": "CRIL: Continual Robot Imitation Learning via Generative and Prediction Model", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-06-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-07-02", "url": "https://arxiv.org/abs/2106.09422"}
{"aliases": [], "authors": ["Haoyu Dong", "Shijie Liu", "Shi Han", "Zhouyu Fu", "Dongmei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "23e0ff317a7b0e4cc8a5535d291f158bf3b184b1bd5dea5a01962f0a1d0cc0bb", "submitted_at": "2021-06-25T08:41:01Z", "updated_at": "2021-06-25T08:41:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2106.13500v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2106.13500", "identifiers": {"arxiv": "2106.13500"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2106.13500"}], "name": "TableSense: Spreadsheet Table Detection with Convolutional Neural Networks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-06-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-06-25", "url": "https://arxiv.org/abs/2106.13500"}
{"aliases": [], "authors": ["Mushahid Khan", "Andriy Miranskyy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a497b59ad6829d4d2a349c3cbfe49a20abb0a198d88f7c27aac048f3fa06ac3", "submitted_at": "2021-06-30T16:09:29Z", "updated_at": "2021-06-30T16:09:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2106.16173v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:05Z", "role": "primary", "sha256": "bb3d2692c294743a8d06d5065cc84b27b229a981c2c650ff20a2ec038797a0e3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=1000&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2106.16173", "identifiers": {"arxiv": "2106.16173"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2106.16173"}], "name": "String Comparison on a Quantum Computer Using Hamming Distance", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-06-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-06-30", "url": "https://arxiv.org/abs/2106.16173"}
{"aliases": [], "authors": ["Mark Chen", "Jerry Tworek", "Heewoo Jun", "Qiming Yuan", "Henrique Ponde de Oliveira Pinto", "Jared Kaplan", "Harri Edwards", "Yuri Burda", "Nicholas Joseph", "Greg Brockman", "Alex Ray", "Raul Puri", "Gretchen Krueger", "Michael Petrov", "Heidy Khlaaf", "Girish Sastry", "Pamela Mishkin", "Brooke Chan", "Scott Gray", "Nick Ryder", "Mikhail Pavlov", "Alethea Power", "Lukasz Kaiser", "Mohammad Bavarian", "Clemens Winter", "Philippe Tillet", "Felipe Petroski Such", "Dave Cummings", "Matthias Plappert", "Fotios Chantzis", "Elizabeth Barnes", "Ariel Herbert-Voss", "William Hebgen Guss", "Alex Nichol", "Alex Paino", "Nikolas Tezak", "Jie Tang", "Igor Babuschkin", "Suchir Balaji", "Shantanu Jain", "William Saunders", "Christopher Hesse", "Andrew N. Carr", "Jan Leike", "Josh Achiam", "Vedant Misra", "Evan Morikawa", "Alec Radford", "Matthew Knight", "Miles Brundage", "Mira Murati", "Katie Mayer", "Peter Welinder", "Bob McGrew", "Dario Amodei", "Sam McCandlish", "Ilya Sutskever", "Wojciech Zaremba"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fdf151f8931c537baf18316784a31cecb4ed2b3c02e7aa62329079db995b4aed", "submitted_at": "2021-07-07T17:41:24Z", "updated_at": "2021-07-14T17:16:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2107.03374v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2107.03374", "identifiers": {"arxiv": "2107.03374"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2107.03374"}], "name": "Evaluating Large Language Models Trained on Code", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-07-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-07-14", "url": "https://arxiv.org/abs/2107.03374"}
{"aliases": [], "authors": ["Yatong Chen", "Zeyu Tang", "Kun Zhang", "Yang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "16d2a67c0ba0e2bea59f7c2ebefbfff480d935da71ba1dc387f14482dd4e52a2", "submitted_at": "2021-07-13T08:21:37Z", "updated_at": "2023-06-01T16:07:11Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2107.05911v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2107.05911", "identifiers": {"arxiv": "2107.05911"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2107.05911"}], "name": "Model Transferability With Responsive Decision Subjects", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-07-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-06-01", "url": "https://arxiv.org/abs/2107.05911"}
{"aliases": [], "authors": ["Chongyang Bai", "Xiaoxue Zang", "Ying Xu", "Srinivas Sunkara", "Abhinav Rastogi", "Jindong Chen", "Blaise Aguera y Arcas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a4f582015ccaf5bbe9be4b7ab925b2bbf6bd8451a4e63d89e1fe38363e26b31", "submitted_at": "2021-07-29T03:51:36Z", "updated_at": "2021-08-10T08:55:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2107.13731v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2107.13731", "identifiers": {"arxiv": "2107.13731"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2107.13731"}], "name": "UIBert: Learning Generic Multimodal Representations for UI Understanding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2021-07-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Learns shared image, text, and structural UI representations through self-supervised tasks based on relationships among interface components.", "topics": ["screen-understanding"], "updated_at": "2021-08-10", "url": "https://arxiv.org/abs/2107.13731"}
{"aliases": [], "authors": ["Armin Galetzka", "Dimitrios Loukrezis", "Herbert De Gersem"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "32967e61b255926860ad60ac9a82f22e1d1002ce4a71caaa0df862e356a50fc3", "submitted_at": "2021-07-30T12:57:36Z", "updated_at": "2021-12-01T07:21:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2107.14604v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2107.14604", "identifiers": {"arxiv": "2107.14604"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2107.14604"}], "name": "Three-Dimensional Data-Driven Magnetostatic Field Computation using Real-World Measurement Data", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-07-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-12-01", "url": "https://arxiv.org/abs/2107.14604"}
{"aliases": [], "authors": ["Mohammad Goudarzi", "Qifan Deng", "Rajkumar Buyya"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc2cd57d488b6f7575f3a740590371a200bb1c21d73029adb0126af7c53fb4a3", "submitted_at": "2021-08-02T01:55:41Z", "updated_at": "2021-08-02T01:55:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2108.00591v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2108.00591", "identifiers": {"arxiv": "2108.00591"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2108.00591"}], "name": "Resource Management in Edge and Fog Computing using FogBus2 Framework", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-08-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-08-02", "url": "https://arxiv.org/abs/2108.00591"}
{"aliases": [], "authors": ["Kirill Rivkin", "Michael Montemorra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d62a6c19eb8633f0e14e6d8b195398a2696258e8f14bd15da1a3fe37651bce1d", "submitted_at": "2021-08-02T14:00:17Z", "updated_at": "2021-08-02T14:00:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2108.00909v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2108.00909", "identifiers": {"arxiv": "2108.00909"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2108.00909"}], "name": "Spin Wave Computing using pre-recorded magnetization patterns", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-08-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-08-02", "url": "https://arxiv.org/abs/2108.00909"}
{"aliases": [], "authors": ["Ajay Mandlekar", "Danfei Xu", "Josiah Wong", "Soroush Nasiriany", "Chen Wang", "Rohun Kulkarni", "Li Fei-Fei", "Silvio Savarese", "Yuke Zhu", "Roberto Martín-Martín"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a71938d5b3fa6612a4f5f7a812bb68dcf76835c7364e8b4a683d75a7fc53c651", "submitted_at": "2021-08-06T20:48:30Z", "updated_at": "2021-09-25T00:37:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2108.03298v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2108.03298", "identifiers": {"arxiv": "2108.03298"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2108.03298"}], "name": "What Matters in Learning from Offline Human Demonstrations for Robot Manipulation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-08-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-09-25", "url": "https://arxiv.org/abs/2108.03298"}
{"aliases": [], "authors": ["Bryan Wang", "Gang Li", "Xin Zhou", "Zhourong Chen", "Tovi Grossman", "Yang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9cf8d8521e643e1a888e706dff4e5106f178061d9af4ed2c2ba5ae42f194785a", "submitted_at": "2021-08-07T03:01:23Z", "updated_at": "2021-08-07T03:01:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2108.03353v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2108.03353", "identifiers": {"arxiv": "2108.03353"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2108.03353"}], "name": "Screen2Words: Automatic Mobile UI Summarization with Multimodal Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2021-08-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies mobile-screen summarization using visual, textual, and structural UI information, with a dataset of human-written descriptions.", "topics": ["screen-understanding"], "updated_at": "2021-08-07", "url": "https://arxiv.org/abs/2108.03353"}
{"aliases": [], "authors": ["Davide Rigoni", "Luciano Serafini", "Alessandro Sperduti"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b08e1479450e2a0252346074060dc83c277263c083417305b07e21273c951d20", "submitted_at": "2021-08-11T16:26:54Z", "updated_at": "2022-02-02T10:57:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2108.05308v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2108.05308", "identifiers": {"arxiv": "2108.05308"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2108.05308"}], "name": "A Better Loss for Visual-Textual Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-08-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-02-02", "url": "https://arxiv.org/abs/2108.05308"}
{"aliases": [], "authors": ["Vaibhav Kumar", "Muhammad Fainan Hanif", "Markku Juntti", "Le-Nam Tran"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c7ae4975b04c3638b7775475e837f9888af5854256f8311340af96bcd43c1f38", "submitted_at": "2021-09-02T22:52:00Z", "updated_at": "2023-10-12T21:12:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2109.01239v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2109.01239", "identifiers": {"arxiv": "2109.01239"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2109.01239"}], "name": "A Max-Min Task Offloading Algorithm for Mobile Edge Computing Using Non-Orthogonal Multiple Access", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-09-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-12", "url": "https://arxiv.org/abs/2109.01239"}
{"aliases": [], "authors": ["Cong Wang", "Shi-Jun Liang", "Chen-Yu Wang", "Zai-Zheng Yang", "Yingmeng Ge", "Chen Pan", "Xi Shen", "Wei Wei", "Yichen Zhao", "Zaichen Zhang", "Bin Cheng", "Chuan Zhang", "Feng Miao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9666a8537233c1bf7f329589e394cc34f21609dd06fb535ac04f26345b9ea0b9", "submitted_at": "2021-09-16T13:37:31Z", "updated_at": "2021-09-16T13:37:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2109.07976v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2109.07976", "identifiers": {"arxiv": "2109.07976"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2109.07976"}], "name": "Scalable massively parallel computing using continuous-time data representation in nanoscale crossbar array", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-09-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-09-16", "url": "https://arxiv.org/abs/2109.07976"}
{"aliases": [], "authors": ["Srikrishna Omkar", "Seok-Hyung Lee", "Yong Siah Teo", "Seung-Woo Lee", "Hyunseok Jeong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "953639b0817913b9fb328a77b7d9b2c8315a496b38d2822290c4451412ee72ee", "submitted_at": "2021-09-25T05:48:40Z", "updated_at": "2022-08-01T14:50:28Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2109.12280v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2109.12280", "identifiers": {"arxiv": "2109.12280"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2109.12280"}], "name": "All-photonic architectural roadmap for scalable quantum computing using Greenberger-Horne-Zeilinger states", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-09-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-08-01", "url": "https://arxiv.org/abs/2109.12280"}
{"aliases": [], "authors": ["Isiah Zaplana", "Hugo Hadfield", "Joan Lasenby"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc2c2a63f4d1a21cc32e5ae59412424842a05e4bc16408d17773f2a996b09a58", "submitted_at": "2021-09-26T00:58:42Z", "updated_at": "2021-09-26T00:58:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2109.12470v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2109.12470", "identifiers": {"arxiv": "2109.12470"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2109.12470"}], "name": "Singularities of serial robots: Identification and distance computation using geometric algebra", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-09-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-09-26", "url": "https://arxiv.org/abs/2109.12470"}
{"aliases": [], "authors": ["Yong Zhang", "Konglong Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f4474b69c78cf9e057c5bebec9c09a5a5d31d756b99d2d610e578fc461334b8", "submitted_at": "2021-09-24T12:04:27Z", "updated_at": "2021-09-24T12:04:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2109.13639v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2109.13639", "identifiers": {"arxiv": "2109.13639"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2109.13639"}], "name": "Quantum Computation Using Action Variables", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-09-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-09-24", "url": "https://arxiv.org/abs/2109.13639"}
{"aliases": [], "authors": ["Akira Imakura", "Keiichi Morikuni", "Akitoshi Takayasu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "979e942932bd25fff3f0b97bb6d06a2ace61e02c2b40de4823b794a41c384fd6", "submitted_at": "2021-10-05T05:04:27Z", "updated_at": "2022-10-25T06:19:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2110.01822v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2110.01822", "identifiers": {"arxiv": "2110.01822"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2110.01822"}], "name": "Verified eigenvalue and eigenvector computations using complex moments and the Rayleigh$\\unicode{x2013}$Ritz procedure for generalized Hermitian eigenvalue problems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-10-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-10-25", "url": "https://arxiv.org/abs/2110.01822"}
{"aliases": [], "authors": ["Tim Baer", "Raghavendra Kanakagiri", "Edgar Solomonik"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6877f400826214748e3af90436ff001f3969db6ee15746de11e8f8b49efb8f5c", "submitted_at": "2021-10-10T17:46:27Z", "updated_at": "2021-12-15T03:36:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2110.04865v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2110.04865", "identifiers": {"arxiv": "2110.04865"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2110.04865"}], "name": "Parallel Minimum Spanning Forest Computation using Sparse Matrix Kernels", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-10-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-12-15", "url": "https://arxiv.org/abs/2110.04865"}
{"aliases": [], "authors": ["Victor Zhong", "Austin W. Hanjie", "Sida I. Wang", "Karthik Narasimhan", "Luke Zettlemoyer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5799d04536a7520114ff02a62614656ebea39d8646dd0367925b5ae72a96b10d", "submitted_at": "2021-10-20T17:02:06Z", "updated_at": "2022-01-24T19:16:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2110.10661v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2110.10661", "identifiers": {"arxiv": "2110.10661"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2110.10661"}], "name": "SILG: The Multi-environment Symbolic Interactive Language Grounding Benchmark", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-01-24", "url": "https://arxiv.org/abs/2110.10661"}
{"aliases": [], "authors": ["Lawrence Z. Cohen", "Isaac H. Kim", "Stephen D. Bartlett", "Benjamin J. Brown"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0110623b83cc7684e6aa46b4cbf211c246cde0c317e01276b39031285af519f", "submitted_at": "2021-10-20T21:49:48Z", "updated_at": "2022-05-22T19:19:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2110.10794v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2110.10794", "identifiers": {"arxiv": "2110.10794"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2110.10794"}], "name": "Low-overhead fault-tolerant quantum computing using long-range connectivity", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-05-22", "url": "https://arxiv.org/abs/2110.10794"}
{"aliases": [], "authors": ["Saeed Ahmed Khan", "Fangjun Hu", "Gerasimos Angelatos", "Hakan E. Türeci"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da12f225a80a6784a1d57f30f90e39712f33385367cf7e4bb2eb729975eb46c1", "submitted_at": "2021-10-26T16:46:14Z", "updated_at": "2021-11-05T17:38:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2110.13849v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2110.13849", "identifiers": {"arxiv": "2110.13849"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2110.13849"}], "name": "Physical reservoir computing using finitely-sampled quantum systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-10-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-11-05", "url": "https://arxiv.org/abs/2110.13849"}
{"aliases": [], "authors": ["Duc M. Tran", "Duy V. Nguyen", "Le Bin Ho", "Hung Q. Nguyen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "96437844b739d0b6e0a90e0df60ad41249f5cb6b9eb6015cc3bb2da189d8947e", "submitted_at": "2021-11-02T15:52:49Z", "updated_at": "2022-02-08T04:05:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2111.02896v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2111.02896", "identifiers": {"arxiv": "2111.02896"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2111.02896"}], "name": "Experimenting quantum phenomena on NISQ computers using high level quantum programming", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-11-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-02-08", "url": "https://arxiv.org/abs/2111.02896"}
{"aliases": [], "authors": ["Max Cubillos", "Edwin Jimenez"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "674e93fe46a816098c2542ce6f3c23a0ef4dc7e6444ec62a5cf16829a248da08", "submitted_at": "2021-11-12T21:36:48Z", "updated_at": "2021-11-12T21:36:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2111.06951v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2111.06951", "identifiers": {"arxiv": "2111.06951"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2111.06951"}], "name": "Diffraction integral computation using sinc approximation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-11-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-11-12", "url": "https://arxiv.org/abs/2111.06951"}
{"aliases": [], "authors": ["Rodrigo Araiza Bravo", "Khadijeh Najafi", "Xun Gao", "Susanne F. Yelin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5973800b41bd008bcf11962ad125068908931d45ff15b647913d330ba5f94633", "submitted_at": "2021-11-22T02:45:18Z", "updated_at": "2022-07-20T18:23:34Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2111.10956v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2111.10956", "identifiers": {"arxiv": "2111.10956"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2111.10956"}], "name": "Quantum reservoir computing using arrays of Rydberg atoms", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-11-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-07-20", "url": "https://arxiv.org/abs/2111.10956"}
{"aliases": [], "authors": ["Qian Yu", "Alberto M. Alonso", "Jackie Caminiti", "Kristin M. Beck", "R. Tyler Sutherland", "Dietrich Leibfried", "Kayla J. Rodriguez", "Madhav Dhital", "Boerge Hemmerling", "Hartmut Häffner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "706d921629bedf0a1349c61933e754b24bb984acb8103f6d9d97e015bb73ad2b", "submitted_at": "2021-12-07T23:10:19Z", "updated_at": "2021-12-07T23:10:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2112.04034v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2112.04034", "identifiers": {"arxiv": "2112.04034"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2112.04034"}], "name": "Feasibility study of quantum computing using trapped electrons", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-12-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-12-07", "url": "https://arxiv.org/abs/2112.04034"}
{"aliases": [], "authors": ["Lorenz Baumgarten", "Stefan Bornholdt"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "16ef9da545ae180d03fbe9af1d13bc74c684fe3127f09b5ad8bd3b00353edf73", "submitted_at": "2021-12-10T14:12:21Z", "updated_at": "2021-12-10T14:12:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2112.05558v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2112.05558", "identifiers": {"arxiv": "2112.05558"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2112.05558"}], "name": "Universal computation using localized limit-cycle attractors in neural networks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-12-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-12-10", "url": "https://arxiv.org/abs/2112.05558"}
{"aliases": [], "authors": ["Yang Li", "Gang Li", "Xin Zhou", "Mostafa Dehghani", "Alexey Gritsenko"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e88e6e2659df232174d9fc71a2fac61a021941f096355f8ea3cd4c618c223d3", "submitted_at": "2021-12-10T17:37:26Z", "updated_at": "2021-12-10T17:37:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2112.05692v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2112.05692", "identifiers": {"arxiv": "2112.05692"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2112.05692"}], "name": "VUT: Versatile UI Transformer for Multi-Modal Multi-Task User Interface Modeling", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2021-12-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "The Versatile UI Transformer jointly models interface screenshots, structure, and language for five UI tasks. Its encoder supports element detection when structural metadata is absent, while an autoregressive component handles UI question answering and command grounding.", "topics": ["screen-understanding", "grounding", "efficiency"], "updated_at": "2021-12-10", "url": "https://arxiv.org/abs/2112.05692"}
{"aliases": [], "authors": ["Anas Skalli", "Joshua Robertson", "Dafydd Owen-Newns", "Matej Hejda", "Xavier Porte", "Stephan Reitzenstein", "Antonio Hurtado", "D. Brunner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f85e7b074d23fe141edf952cdc4850565f813fd6f0d411979c7366073f8bb7c3", "submitted_at": "2021-12-15T12:45:39Z", "updated_at": "2021-12-15T12:45:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2112.08086v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2112.08086", "identifiers": {"arxiv": "2112.08086"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2112.08086"}], "name": "Photonic neuromorphic computing using vertical cavity semiconductor lasers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-12-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-12-15", "url": "https://arxiv.org/abs/2112.08086"}
{"aliases": [], "authors": ["Moein Owhadi-Kareshk", "Pierre Boulanger"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b51674fae971f4d30255e6167143d2793c6e0270914f79ce7b46634623eaedb0", "submitted_at": "2021-12-16T16:30:23Z", "updated_at": "2021-12-16T16:30:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2112.08998v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2112.08998", "identifiers": {"arxiv": "2112.08998"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2112.08998"}], "name": "Portfolio Optimization on Classical and Quantum Computers Using PortFawn", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-12-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-12-16", "url": "https://arxiv.org/abs/2112.08998"}
{"aliases": [], "authors": ["Reiichiro Nakano", "Jacob Hilton", "Suchir Balaji", "Jeff Wu", "Long Ouyang", "Christina Kim", "Christopher Hesse", "Shantanu Jain", "Vineet Kosaraju", "William Saunders", "Xu Jiang", "Karl Cobbe", "Tyna Eloundou", "Gretchen Krueger", "Kevin Button", "Matthew Knight", "Benjamin Chess", "John Schulman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a9122bd2c771071f93250e9e330b1f923f1e50c0d829cc3144be43c6b575195", "submitted_at": "2021-12-17T05:43:43Z", "updated_at": "2022-06-01T19:08:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2112.09332v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2112.09332", "identifiers": {"arxiv": "2112.09332"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2112.09332"}], "name": "WebGPT: Browser-assisted question-answering with human feedback", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-12-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-06-01", "url": "https://arxiv.org/abs/2112.09332"}
{"aliases": [], "authors": ["Timothy Proctor", "Stefan Seritan", "Kenneth Rudinger", "Erik Nielsen", "Robin Blume-Kohout", "Kevin Young"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "100ab5b501d1ed147dccc539239ebbbd8bdc0f7d4c9f8de4f0ef1e4fc8c8bbff", "submitted_at": "2021-12-18T06:04:22Z", "updated_at": "2022-10-10T15:36:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2112.09853v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2112.09853", "identifiers": {"arxiv": "2112.09853"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2112.09853"}], "name": "Scalable randomized benchmarking of quantum computers using mirror circuits", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-12-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-10-10", "url": "https://arxiv.org/abs/2112.09853"}
{"aliases": [], "authors": ["Riad Ladjel", "Nicolas Anciaux", "Aurélien Bellet", "Guillaume Scerri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0c554f0021e24d4843f94ce483b7a94f4c64843872b9bcf3710ca515616b956", "submitted_at": "2021-12-23T08:30:17Z", "updated_at": "2021-12-23T08:30:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2112.12411v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2112.12411", "identifiers": {"arxiv": "2112.12411"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2112.12411"}], "name": "Mitigating Leakage from Data Dependent Communications in Decentralized Computing using Differential Privacy", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-12-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2021-12-23", "url": "https://arxiv.org/abs/2112.12411"}
{"aliases": [], "authors": ["Alain Denzler", "Rahul Bera", "Nastaran Hajinazar", "Gagandeep Singh", "Geraldo F. Oliveira", "Juan Gómez-Luna", "Onur Mutlu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "065a3310e52d540659bf73406838c93eabd6109b35cdb83b398835a5c9ad1149", "submitted_at": "2021-12-28T16:45:56Z", "updated_at": "2023-09-05T20:13:46Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2112.14216v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2112.14216", "identifiers": {"arxiv": "2112.14216"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2112.14216"}], "name": "Casper: Accelerating Stencil Computation using Near-cache Processing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2021-12-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-09-05", "url": "https://arxiv.org/abs/2112.14216"}
{"aliases": [], "authors": ["Yu. I. Bogdanov", "N. A. Bogdanova", "D. V. Fastovets", "V. F. Lukichev"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e91ccc695c4d61c1e569a399864a3310c87d85d6b73c5b4159addcc86365c18", "submitted_at": "2022-01-09T19:02:43Z", "updated_at": "2022-01-09T19:02:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2201.03068v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2201.03068", "identifiers": {"arxiv": "2201.03068"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2201.03068"}], "name": "Study of the effect of quantum noise on the accuracy of the Schrodinger equation simulation on a quantum computer using the Zalka-Wiesner method", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-01-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-01-09", "url": "https://arxiv.org/abs/2201.03068"}
{"aliases": [], "authors": ["Gang Li", "Gilles Baechler", "Manuel Tragut", "Yang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d971ee06cafd46fe27a01d40bc16e3604e1d153c38ee92be970328d9935b0708", "submitted_at": "2022-01-11T17:52:40Z", "updated_at": "2022-01-13T17:53:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2201.04100v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:47Z", "role": "primary", "sha256": "30f2d08b6bf51489cf780365753d9e9e945156951014ae76501b5796eaacf83c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2201.04100v2"}], "id": "arxiv:2201.04100", "identifiers": {"arxiv": "2201.04100"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2201.04100"}], "name": "Learning to Denoise Raw Mobile UI Layouts for Improving Datasets at Scale", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2022-01-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CLAY, a pipeline that repairs noisy mobile UI layouts by removing nodes without valid visual counterparts and assigning semantic element types, supported by human-annotated layouts for evaluation.", "topics": ["screen-understanding", "data-generation"], "updated_at": "2022-01-13", "url": "https://arxiv.org/abs/2201.04100"}
{"aliases": [], "authors": ["Sandeep Kumar Patel", "Avtar Singh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5180b6e42f65d68297a82fab7ac68cb52f2598b8699d12228f24339166102157", "submitted_at": "2022-01-23T12:20:21Z", "updated_at": "2022-01-23T12:20:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2201.09242v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2201.09242", "identifiers": {"arxiv": "2201.09242"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2201.09242"}], "name": "Task Scheduling in Cloud Computing Using Hybrid Meta-heuristic: A Review", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-01-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-01-23", "url": "https://arxiv.org/abs/2201.09242"}
{"aliases": [], "authors": ["Kazuhiro Seki", "Yuichi Otsuka", "Seiji Yunoki"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8bcd4e5acf5b762136ad9ce65f0397b22169449aa0ccd10c5175166795c47ecc", "submitted_at": "2022-01-27T08:57:04Z", "updated_at": "2022-04-14T01:07:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2201.11381v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2201.11381", "identifiers": {"arxiv": "2201.11381"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2201.11381"}], "name": "Gutzwiller wave function on a quantum computer using a discrete Hubbard-Stratonovich transformation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-01-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-04-14", "url": "https://arxiv.org/abs/2201.11381"}
{"aliases": [], "authors": ["Carlos Mougan", "Dan Saattrup Nielsen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e10c22aa4cfc6d4e3f9221b93cf9737f942c39213eaeca891d31187fa698e4c9", "submitted_at": "2022-01-27T17:23:04Z", "updated_at": "2022-11-22T11:54:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2201.11676v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2201.11676", "identifiers": {"arxiv": "2201.11676"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2201.11676"}], "name": "Monitoring Model Deterioration with Explainable Uncertainty Estimation via Non-parametric Bootstrap", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-01-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-11-22", "url": "https://arxiv.org/abs/2201.11676"}
{"aliases": [], "authors": ["Mohammad Reza Jokar", "Richard Rines", "Ghasem Pasandi", "Haolin Cong", "Adam Holmes", "Yunong Shi", "Massoud Pedram", "Frederic T. Chong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f244ab0734c6e8969e5e09ad8138f4de6854004fa2ecb1a560eb26d672b2337d", "submitted_at": "2022-02-03T04:52:14Z", "updated_at": "2022-02-03T04:52:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2202.01407v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2202.01407", "identifiers": {"arxiv": "2202.01407"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2202.01407"}], "name": "DigiQ: A Scalable Digital Controller for Quantum Computers Using SFQ Logic", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-02-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-02-03", "url": "https://arxiv.org/abs/2202.01407"}
{"aliases": [], "authors": ["Andrea Burns", "Deniz Arsan", "Sanjna Agrawal", "Ranjitha Kumar", "Kate Saenko", "Bryan A. Plummer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1687369e31b0ad3632381d4a46b1d555c2e2ed1d6504e9c1b63006dd29f3445e", "submitted_at": "2022-02-04T18:51:50Z", "updated_at": "2022-08-15T00:24:24Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2202.02312v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2202.02312", "identifiers": {"arxiv": "2202.02312"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2202.02312"}], "name": "A Dataset for Interactive Vision-Language Navigation with Unknown Command Feasibility", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2022-02-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MoTIF mobile-app interactions with feasible and infeasible commands, feasibility explanations, and follow-up questions for ambiguous requests.", "topics": ["mobile-navigation", "human-in-the-loop", "evaluation"], "updated_at": "2022-08-15", "url": "https://arxiv.org/abs/2202.02312"}
{"aliases": [], "authors": ["Tommaso Sorgente", "Silvia Biasotti", "Michela Spagnuolo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "39fba53b7c06d069f6da3c5a66e6215c8a9637e31dcf299ed2d57986b30f2860", "submitted_at": "2022-02-14T11:25:14Z", "updated_at": "2022-02-14T11:25:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2202.06625v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2202.06625", "identifiers": {"arxiv": "2202.06625"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2202.06625"}], "name": "Polyhedron Kernel Computation Using a Geometric Approach", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-02-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-02-14", "url": "https://arxiv.org/abs/2202.06625"}
{"aliases": [], "authors": ["Peter C Humphreys", "David Raposo", "Toby Pohlen", "Gregory Thornton", "Rachita Chhaparia", "Alistair Muldal", "Josh Abramson", "Petko Georgiev", "Alex Goldin", "Adam Santoro", "Timothy Lillicrap"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "94129847ac85f4a8aeab74907cc68a31c435b078ab950695d9328b6c3896b46d", "submitted_at": "2022-02-16T15:23:46Z", "updated_at": "2022-11-11T13:45:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2202.08137v2"}, "publication": {"authors": ["Peter Humphreys", "David Raposo", "Toby Pohlen", "Gregory Thornton", "Rachita Chhaparia", "Alistair Muldal", "Josh Abramson", "Petko Georgiev", "Adam Santoro", "Timothy Lillicrap"], "date_basis": "Explicit publisher metadata: 2022/06/28; precision=day.", "date_parts": [2022, 6, 28], "doi": null, "event_dates": "17--23 Jul 2022", "evidence_urls": ["https://proceedings.mlr.press/v162/humphreys22a.html", "https://proceedings.mlr.press/v162/humphreys22a/humphreys22a.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The formal ICML PDF lists 10 authors beginning Peter Humphreys and uses Toby Pohlen. The landing lists Peter C Humphreys and Tobias Pohlen. The retained arXiv snapshot lists 11 authors, including Alex Goldin. These are preserved as distinct version-specific author lists; the publication field uses only the formal PDF list. The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher issue date is 2022-06-28; the July 17–23 conference range is an event date, not a substitute publication day. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "9466-9482", "pdf_url": "https://proceedings.mlr.press/v162/humphreys22a/humphreys22a.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "A Data-Driven Approach for Learning to Control Computers", "type": "conference-paper", "url": "https://proceedings.mlr.press/v162/humphreys22a.html", "venue": "Proceedings of the 39th International Conference on Machine Learning", "volume": "162"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:47+00:00", "role": "primary", "sha256": "b0ac6ec4bf3aad3773c53a0fffe8688a6287f25f28faf3743603ec2c1ae08a8b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.series", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v162/humphreys22a.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:46+00:00", "role": "primary", "sha256": "ba2a9a0a5e46aa5f34cfc720ac08049253917407e2fbd4a8cdbd2c8ca3ec8455", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://proceedings.mlr.press/v162/humphreys22a/humphreys22a.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2202.08137", "identifiers": {"arxiv": "2202.08137"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2202.08137"}, {"rel": "publication", "url": "https://proceedings.mlr.press/v162/humphreys22a.html"}, {"rel": "publication_pdf", "url": "https://proceedings.mlr.press/v162/humphreys22a/humphreys22a.pdf"}], "name": "A data-driven approach for learning to control computers", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2022-02-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines reinforcement learning with behavioral priors from human interactions to learn computer control through keyboard and mouse actions.", "topics": ["reinforcement-learning", "automation"], "updated_at": "2022-11-11", "url": "https://arxiv.org/abs/2202.08137"}
{"aliases": [], "authors": ["Bolin Zhang", "Yu Liu", "Tianqi Gao", "Deming Zhang", "Weisheng Zhao", "Lang Zeng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c7aa8a13e656ce3060de53fa719e3b5d9d6cbcf50a6c273dda9e1e3bde576844", "submitted_at": "2022-02-22T02:47:08Z", "updated_at": "2022-02-22T02:47:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2202.10634v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2202.10634", "identifiers": {"arxiv": "2202.10634"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2202.10634"}], "name": "Time Division Multiplexing Ising Computer Using Single Tunable True Random Number Generator Based on Spin Torque Nano-Oscillator", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-02-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-02-22", "url": "https://arxiv.org/abs/2202.10634"}
{"aliases": [], "authors": ["Filipa C. R. Peres", "Ernesto F. Galvão"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c0e60df1c4b4e057d5f12dd4fd03a55e221ec5d7f7c12416290fbb60459fb70", "submitted_at": "2022-03-03T16:01:55Z", "updated_at": "2023-09-28T13:38:16Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2203.01789v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2203.01789", "identifiers": {"arxiv": "2203.01789"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2203.01789"}], "name": "Quantum circuit compilation and hybrid computation using Pauli-based computation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-03-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-09-28", "url": "https://arxiv.org/abs/2203.01789"}
{"aliases": [], "authors": ["Hsuan Hsiao", "Vincent Lee", "Brandon Reagen", "Armin Alaghi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0c7420280bb8f7deea17793c576a7eeed073ee1fcb7a157a60b960db27390ac", "submitted_at": "2022-03-04T19:53:10Z", "updated_at": "2022-03-04T19:53:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2203.02547v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2203.02547", "identifiers": {"arxiv": "2203.02547"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2203.02547"}], "name": "Homomorphically Encrypted Computation using Stochastic Encodings", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-03-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-03-04", "url": "https://arxiv.org/abs/2203.02547"}
{"aliases": [], "authors": ["Ye-Hong Chen", "Adam Miranowicz", "Xi Chen", "Yan Xia", "Franco Nori"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5dde2a79c532ede977bdc5c9bf5d98074c32528b800540d944717498d8b267f9", "submitted_at": "2022-03-14T03:06:15Z", "updated_at": "2022-12-21T10:55:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2203.06831v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2203.06831", "identifiers": {"arxiv": "2203.06831"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2203.06831"}], "name": "Enhanced-Fidelity Ultrafast Geometric Quantum Computation Using Strong Classical Drives", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-03-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-12-21", "url": "https://arxiv.org/abs/2203.06831"}
{"aliases": [], "authors": ["Genji Fujii"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "20caa27a6993167cdd1c1c9be912a91270ec28511ef0e670b0451a463f75e7a0", "submitted_at": "2022-03-16T23:51:04Z", "updated_at": "2025-09-30T06:19:50Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2203.08993v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2203.08993", "identifiers": {"arxiv": "2203.08993"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2203.08993"}], "name": "Measurement-based quantum computation using two-component BECs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-03-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-30", "url": "https://arxiv.org/abs/2203.08993"}
{"aliases": [], "authors": ["Xiaosu Zhu", "Jingkuan Song", "Lianli Gao", "Feng Zheng", "Heng Tao Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5337b307d37f1b490f65b700f3949f617e2a4c2c17a1d8aff4dfe1ab03163f13", "submitted_at": "2022-03-21T11:44:17Z", "updated_at": "2022-03-21T11:44:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2203.10897v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2203.10897", "identifiers": {"arxiv": "2203.10897"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2203.10897"}], "name": "Unified Multivariate Gaussian Mixture for Efficient Neural Image Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-03-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-03-21", "url": "https://arxiv.org/abs/2203.10897"}
{"aliases": [], "authors": ["K. Sozos", "A. Bogris", "P. Bienstman", "G. Sarantoglou", "S. Deligiannidis", "C. Mesaritakis"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2e528a39d84abdbc6500d52e6578d82b78e706434d9da5587db3320eb4f8fe98", "submitted_at": "2022-03-29T09:13:00Z", "updated_at": "2022-03-29T09:13:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2203.15807v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2203.15807", "identifiers": {"arxiv": "2203.15807"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2203.15807"}], "name": "High Speed Photonic Neuromorphic Computing Using Recurrent Optical Spectrum Slicing Neural Networks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-03-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-03-29", "url": "https://arxiv.org/abs/2203.15807"}
{"aliases": [], "authors": ["Michael Ahn", "Anthony Brohan", "Noah Brown", "Yevgen Chebotar", "Omar Cortes", "Byron David", "Chelsea Finn", "Chuyuan Fu", "Keerthana Gopalakrishnan", "Karol Hausman", "Alex Herzog", "Daniel Ho", "Jasmine Hsu", "Julian Ibarz", "Brian Ichter", "Alex Irpan", "Eric Jang", "Rosario Jauregui Ruano", "Kyle Jeffrey", "Sally Jesmonth", "Nikhil J Joshi", "Ryan Julian", "Dmitry Kalashnikov", "Yuheng Kuang", "Kuang-Huei Lee", "Sergey Levine", "Yao Lu", "Linda Luu", "Carolina Parada", "Peter Pastor", "Jornell Quiambao", "Kanishka Rao", "Jarek Rettinghouse", "Diego Reyes", "Pierre Sermanet", "Nicolas Sievers", "Clayton Tan", "Alexander Toshev", "Vincent Vanhoucke", "Fei Xia", "Ted Xiao", "Peng Xu", "Sichun Xu", "Mengyuan Yan", "Andy Zeng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccf2ec995c14bd57bb0b111d02a3a1029f72930de8610f54ec3691143b29c0f5", "submitted_at": "2022-04-04T17:57:11Z", "updated_at": "2022-08-16T16:06:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2204.01691v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2204.01691", "identifiers": {"arxiv": "2204.01691"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2204.01691"}], "name": "Do As I Can, Not As I Say: Grounding Language in Robotic Affordances", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-04-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-08-16", "url": "https://arxiv.org/abs/2204.01691"}
{"aliases": [], "authors": ["Murat Işık", "Ankita Paul", "M. Lakshmi Varshika", "Anup Das"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03fa41fc3d96d0f4982b1c08ea08aeec80454b5c9810ade38fcce816580de52e", "submitted_at": "2022-04-06T17:15:14Z", "updated_at": "2022-04-06T17:15:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2204.02942v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2204.02942", "identifiers": {"arxiv": "2204.02942"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2204.02942"}], "name": "A Design Methodology for Fault-Tolerant Computing using Astrocyte Neural Networks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-04-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-04-06", "url": "https://arxiv.org/abs/2204.02942"}
{"aliases": [], "authors": ["Lei Xu", "Yangyi Chen", "Ganqu Cui", "Hongcheng Gao", "Zhiyuan Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33c6c9dfe9315c042aae6d441ecdc0ec038a61182f2a9721098fad071763a6a7", "submitted_at": "2022-04-11T16:34:10Z", "updated_at": "2022-04-11T16:34:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2204.05239v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2204.05239", "identifiers": {"arxiv": "2204.05239"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2204.05239"}], "name": "Exploring the Universal Vulnerability of Prompt-based Learning Paradigm", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-04-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-04-11", "url": "https://arxiv.org/abs/2204.05239"}
{"aliases": [], "authors": ["Tomoyuki Yokouchi", "Satoshi Sugimoto", "Bivas Rana", "Shinichiro Seki", "Naoki Ogawa", "Yuki Shiomi", "Shinya Kasai", "Yoshichika Otani"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1dab800d473a203a5ec3f6f3912aacece4c7801a547b24de9fc6e64dfcc7b5f1", "submitted_at": "2022-04-28T02:45:42Z", "updated_at": "2022-04-28T02:45:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2204.13260v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2204.13260", "identifiers": {"arxiv": "2204.13260"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2204.13260"}], "name": "Pattern recognition with neuromorphic computing using magnetic-field induced dynamics of skyrmions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-04-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-04-28", "url": "https://arxiv.org/abs/2204.13260"}
{"aliases": [], "authors": ["Ehud Karpas", "Omri Abend", "Yonatan Belinkov", "Barak Lenz", "Opher Lieber", "Nir Ratner", "Yoav Shoham", "Hofit Bata", "Yoav Levine", "Kevin Leyton-Brown", "Dor Muhlgay", "Noam Rozen", "Erez Schwartz", "Gal Shachaf", "Shai Shalev-Shwartz", "Amnon Shashua", "Moshe Tenenholtz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e9ab52d6f36bdc4b02ff204ea00d0a6a4b4ab8a6987b43f3c8dbad4fe7649ab8", "submitted_at": "2022-05-01T11:01:28Z", "updated_at": "2022-05-01T11:01:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2205.00445v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2205.00445", "identifiers": {"arxiv": "2205.00445"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2205.00445"}], "name": "MRKL Systems: A modular, neuro-symbolic architecture that combines large language models, external knowledge sources and discrete reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-05-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-05-01", "url": "https://arxiv.org/abs/2205.00445"}
{"aliases": [], "authors": ["Mathias Brossard", "Guilhem Bryant", "Basma El Gaabouri", "Xinxin Fan", "Alexandre Ferreira", "Edmund Grimley-Evans", "Christopher Haster", "Evan Johnson", "Derek Miller", "Fan Mo", "Dominic P. Mulligan", "Nick Spinale", "Eric van Hensbergen", "Hugo J. M. Vincent", "Shale Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "683662afceb62c690da787e7d5c84ff559101de07e9de43d49394acfe1ebae9d", "submitted_at": "2022-05-06T15:56:40Z", "updated_at": "2022-05-06T15:56:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2205.03322v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2205.03322", "identifiers": {"arxiv": "2205.03322"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2205.03322"}], "name": "Private delegated computations using strong isolation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-05-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-05-06", "url": "https://arxiv.org/abs/2205.03322"}
{"aliases": [], "authors": ["Artyom M. Grigoryan", "Sos S. Agaian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0220612dcd8ce2433e561ec455e61398dd9cf28572a8a791f247ebb2d6e7aba5", "submitted_at": "2022-05-11T18:53:43Z", "updated_at": "2022-05-11T18:53:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2205.05727v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2205.05727", "identifiers": {"arxiv": "2205.05727"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2205.05727"}], "name": "3-Qubit Circular Quantum Convolution Computation using Fourier Transform with Illustrative Examples", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-05-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-05-11", "url": "https://arxiv.org/abs/2205.05727"}
{"aliases": [], "authors": ["Mengjiao Yang", "Dale Schuurmans", "Pieter Abbeel", "Ofir Nachum"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf2409d6dbd8c1ca857d15c127617a527297bdb680f34de9954883fa82950b10", "submitted_at": "2022-05-22T13:14:09Z", "updated_at": "2022-05-22T13:14:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2205.10816v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2205.10816", "identifiers": {"arxiv": "2205.10816"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2205.10816"}], "name": "Chain of Thought Imitation with Procedure Cloning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-05-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-05-22", "url": "https://arxiv.org/abs/2205.10816"}
{"aliases": [], "authors": ["Liangtai Sun", "Xingyu Chen", "Lu Chen", "Tianle Dai", "Zichen Zhu", "Kai Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbe2a4ece7ef72dfa008740991a5c46d9ab4b9de73f3de02f5dabcee2a5c1aa5", "submitted_at": "2022-05-23T04:05:37Z", "updated_at": "2022-11-24T06:34:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2205.11029v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2205.11029", "identifiers": {"arxiv": "2205.11029"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2205.11029"}], "name": "META-GUI: Towards Multi-modal Conversational Agents on Mobile GUI", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2022-05-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI-based task-oriented dialogue and a dataset pairing mobile-interface operations with conversational responses.", "topics": ["mobile-navigation", "human-in-the-loop"], "updated_at": "2022-11-24", "url": "https://arxiv.org/abs/2205.11029"}
{"aliases": [], "authors": ["Adhitha Dias", "Kirshanthan Sundararajah", "Charitha Saumya", "Milind Kulkarni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1050ed871bd4dda063ad93feb312a4d36fa8975f8a0927e0581b8f9fd440bd95", "submitted_at": "2022-05-23T20:47:42Z", "updated_at": "2022-05-23T20:47:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2205.11622v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2205.11622", "identifiers": {"arxiv": "2205.11622"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2205.11622"}], "name": "SparseLNR: Accelerating Sparse Tensor Computations Using Loop Nest Restructuring", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-05-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-05-23", "url": "https://arxiv.org/abs/2205.11622"}
{"aliases": [], "authors": ["Aaron Parisi", "Yao Zhao", "Noah Fiedel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bc22cd95e8c49cbc444306d3971a3bc839f07246a7c87f88aabfc996b8496e83", "submitted_at": "2022-05-24T17:58:13Z", "updated_at": "2022-05-24T17:58:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2205.12255v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2205.12255", "identifiers": {"arxiv": "2205.12255"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2205.12255"}], "name": "TALM: Tool Augmented Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-05-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-05-24", "url": "https://arxiv.org/abs/2205.12255"}
{"aliases": [], "authors": ["Jun Yan", "Vansh Gupta", "Xiang Ren"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2825ba289924bc5d5ef08693b0e76be70b71f81b7b6a621403050a2a64e95bd5", "submitted_at": "2022-05-25T11:58:38Z", "updated_at": "2023-05-29T17:59:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2205.12700v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2205.12700", "identifiers": {"arxiv": "2205.12700"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2205.12700"}], "name": "BITE: Textual Backdoor Attacks with Iterative Trigger Injection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-05-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-29", "url": "https://arxiv.org/abs/2205.12700"}
{"aliases": [], "authors": ["Li Liu", "Shengping Liu", "Hui Xie", "Fansheng Xiong", "Tengchao Yu", "Mengjuan Xiao", "Lufeng Liu", "Heng Yong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3ab51b75814ac49411b15090a76a1cc6652b9436af9f0a2182459e6523e2bcd9", "submitted_at": "2022-06-05T12:41:06Z", "updated_at": "2022-08-06T10:51:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2206.03864v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2206.03864", "identifiers": {"arxiv": "2206.03864"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2206.03864"}], "name": "Discontinuity Computing using Physics-Informed Neural Network", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-06-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-08-06", "url": "https://arxiv.org/abs/2206.03864"}
{"aliases": [], "authors": ["Aarohi Srivastava", "Abhinav Rastogi", "Abhishek Rao", "Abu Awal Md Shoeb", "Abubakar Abid", "Adam Fisch", "Adam R. Brown", "Adam Santoro", "Aditya Gupta", "Adrià Garriga-Alonso", "Agnieszka Kluska", "Aitor Lewkowycz", "Akshat Agarwal", "Alethea Power", "Alex Ray", "Alex Warstadt", "Alexander W. Kocurek", "Ali Safaya", "Ali Tazarv", "Alice Xiang", "Alicia Parrish", "Allen Nie", "Aman Hussain", "Amanda Askell", "Amanda Dsouza", "Ambrose Slone", "Ameet Rahane", "Anantharaman S. Iyer", "Anders Andreassen", "Andrea Madotto", "Andrea Santilli", "Andreas Stuhlmüller", "Andrew Dai", "Andrew La", "Andrew Lampinen", "Andy Zou", "Angela Jiang", "Angelica Chen", "Anh Vuong", "Animesh Gupta", "Anna Gottardi", "Antonio Norelli", "Anu Venkatesh", "Arash Gholamidavoodi", "Arfa Tabassum", "Arul Menezes", "Arun Kirubarajan", "Asher Mullokandov", "Ashish Sabharwal", "Austin Herrick", "Avia Efrat", "Aykut Erdem", "Ayla Karakaş", "B. Ryan Roberts", "Bao Sheng Loe", "Barret Zoph", "Bartłomiej Bojanowski", "Batuhan Özyurt", "Behnam Hedayatnia", "Behnam Neyshabur", "Benjamin Inden", "Benno Stein", "Berk Ekmekci", "Bill Yuchen Lin", "Blake Howald", "Bryan Orinion", "Cameron Diao", "Cameron Dour", "Catherine Stinson", "Cedrick Argueta", "César Ferri Ramírez", "Chandan Singh", "Charles Rathkopf", "Chenlin Meng", "Chitta Baral", "Chiyu Wu", "Chris Callison-Burch", "Chris Waites", "Christian Voigt", "Christopher D. Manning", "Christopher Potts", "Cindy Ramirez", "Clara E. Rivera", "Clemencia Siro", "Colin Raffel", "Courtney Ashcraft", "Cristina Garbacea", "Damien Sileo", "Dan Garrette", "Dan Hendrycks", "Dan Kilman", "Dan Roth", "Daniel Freeman", "Daniel Khashabi", "Daniel Levy", "Daniel Moseguí González", "Danielle Perszyk", "Danny Hernandez", "Danqi Chen", "Daphne Ippolito", "Dar Gilboa", "David Dohan", "David Drakard", "David Jurgens", "Debajyoti Datta", "Deep Ganguli", "Denis Emelin", "Denis Kleyko", "Deniz Yuret", "Derek Chen", "Derek Tam", "Dieuwke Hupkes", "Diganta Misra", "Dilyar Buzan", "Dimitri Coelho Mollo", "Diyi Yang", "Dong-Ho Lee", "Dylan Schrader", "Ekaterina Shutova", "Ekin Dogus Cubuk", "Elad Segal", "Eleanor Hagerman", "Elizabeth Barnes", "Elizabeth Donoway", "Ellie Pavlick", "Emanuele Rodola", "Emma Lam", "Eric Chu", "Eric Tang", "Erkut Erdem", "Ernie Chang", "Ethan A. Chi", "Ethan Dyer", "Ethan Jerzak", "Ethan Kim", "Eunice Engefu Manyasi", "Evgenii Zheltonozhskii", "Fanyue Xia", "Fatemeh Siar", "Fernando Martínez-Plumed", "Francesca Happé", "Francois Chollet", "Frieda Rong", "Gaurav Mishra", "Genta Indra Winata", "Gerard de Melo", "Germán Kruszewski", "Giambattista Parascandolo", "Giorgio Mariani", "Gloria Wang", "Gonzalo Jaimovitch-López", "Gregor Betz", "Guy Gur-Ari", "Hana Galijasevic", "Hannah Kim", "Hannah Rashkin", "Hannaneh Hajishirzi", "Harsh Mehta", "Hayden Bogar", "Henry Shevlin", "Hinrich Schütze", "Hiromu Yakura", "Hongming Zhang", "Hugh Mee Wong", "Ian Ng", "Isaac Noble", "Jaap Jumelet", "Jack Geissinger", "Jackson Kernion", "Jacob Hilton", "Jaehoon Lee", "Jaime Fernández Fisac", "James B. Simon", "James Koppel", "James Zheng", "James Zou", "Jan Kocoń", "Jana Thompson", "Janelle Wingfield", "Jared Kaplan", "Jarema Radom", "Jascha Sohl-Dickstein", "Jason Phang", "Jason Wei", "Jason Yosinski", "Jekaterina Novikova", "Jelle Bosscher", "Jennifer Marsh", "Jeremy Kim", "Jeroen Taal", "Jesse Engel", "Jesujoba Alabi", "Jiacheng Xu", "Jiaming Song", "Jillian Tang", "Joan Waweru", "John Burden", "John Miller", "John U. Balis", "Jonathan Batchelder", "Jonathan Berant", "Jörg Frohberg", "Jos Rozen", "Jose Hernandez-Orallo", "Joseph Boudeman", "Joseph Guerr", "Joseph Jones", "Joshua B. Tenenbaum", "Joshua S. Rule", "Joyce Chua", "Kamil Kanclerz", "Karen Livescu", "Karl Krauth", "Karthik Gopalakrishnan", "Katerina Ignatyeva", "Katja Markert", "Kaustubh D. Dhole", "Kevin Gimpel", "Kevin Omondi", "Kory Mathewson", "Kristen Chiafullo", "Ksenia Shkaruta", "Kumar Shridhar", "Kyle McDonell", "Kyle Richardson", "Laria Reynolds", "Leo Gao", "Li Zhang", "Liam Dugan", "Lianhui Qin", "Lidia Contreras-Ochando", "Louis-Philippe Morency", "Luca Moschella", "Lucas Lam", "Lucy Noble", "Ludwig Schmidt", "Luheng He", "Luis Oliveros Colón", "Luke Metz", "Lütfi Kerem Şenel", "Maarten Bosma", "Maarten Sap", "Maartje ter Hoeve", "Maheen Farooqi", "Manaal Faruqui", "Mantas Mazeika", "Marco Baturan", "Marco Marelli", "Marco Maru", "Maria Jose Ramírez Quintana", "Marie Tolkiehn", "Mario Giulianelli", "Martha Lewis", "Martin Potthast", "Matthew L. Leavitt", "Matthias Hagen", "Mátyás Schubert", "Medina Orduna Baitemirova", "Melody Arnaud", "Melvin McElrath", "Michael A. Yee", "Michael Cohen", "Michael Gu", "Michael Ivanitskiy", "Michael Starritt", "Michael Strube", "Michał Swędrowski", "Michele Bevilacqua", "Michihiro Yasunaga", "Mihir Kale", "Mike Cain", "Mimee Xu", "Mirac Suzgun", "Mitch Walker", "Mo Tiwari", "Mohit Bansal", "Moin Aminnaseri", "Mor Geva", "Mozhdeh Gheini", "Mukund Varma T", "Nanyun Peng", "Nathan A. Chi", "Nayeon Lee", "Neta Gur-Ari Krakover", "Nicholas Cameron", "Nicholas Roberts", "Nick Doiron", "Nicole Martinez", "Nikita Nangia", "Niklas Deckers", "Niklas Muennighoff", "Nitish Shirish Keskar", "Niveditha S. Iyer", "Noah Constant", "Noah Fiedel", "Nuan Wen", "Oliver Zhang", "Omar Agha", "Omar Elbaghdadi", "Omer Levy", "Owain Evans", "Pablo Antonio Moreno Casares", "Parth Doshi", "Pascale Fung", "Paul Pu Liang", "Paul Vicol", "Pegah Alipoormolabashi", "Peiyuan Liao", "Percy Liang", "Peter Chang", "Peter Eckersley", "Phu Mon Htut", "Pinyu Hwang", "Piotr Miłkowski", "Piyush Patil", "Pouya Pezeshkpour", "Priti Oli", "Qiaozhu Mei", "Qing Lyu", "Qinlang Chen", "Rabin Banjade", "Rachel Etta Rudolph", "Raefer Gabriel", "Rahel Habacker", "Ramon Risco", "Raphaël Millière", "Rhythm Garg", "Richard Barnes", "Rif A. Saurous", "Riku Arakawa", "Robbe Raymaekers", "Robert Frank", "Rohan Sikand", "Roman Novak", "Roman Sitelew", "Ronan LeBras", "Rosanne Liu", "Rowan Jacobs", "Rui Zhang", "Ruslan Salakhutdinov", "Ryan Chi", "Ryan Lee", "Ryan Stovall", "Ryan Teehan", "Rylan Yang", "Sahib Singh", "Saif M. Mohammad", "Sajant Anand", "Sam Dillavou", "Sam Shleifer", "Sam Wiseman", "Samuel Gruetter", "Samuel R. Bowman", "Samuel S. Schoenholz", "Sanghyun Han", "Sanjeev Kwatra", "Sarah A. Rous", "Sarik Ghazarian", "Sayan Ghosh", "Sean Casey", "Sebastian Bischoff", "Sebastian Gehrmann", "Sebastian Schuster", "Sepideh Sadeghi", "Shadi Hamdan", "Sharon Zhou", "Shashank Srivastava", "Sherry Shi", "Shikhar Singh", "Shima Asaadi", "Shixiang Shane Gu", "Shubh Pachchigar", "Shubham Toshniwal", "Shyam Upadhyay", "Shyamolima", "Debnath", "Siamak Shakeri", "Simon Thormeyer", "Simone Melzi", "Siva Reddy", "Sneha Priscilla Makini", "Soo-Hwan Lee", "Spencer Torene", "Sriharsha Hatwar", "Stanislas Dehaene", "Stefan Divic", "Stefano Ermon", "Stella Biderman", "Stephanie Lin", "Stephen Prasad", "Steven T. Piantadosi", "Stuart M. Shieber", "Summer Misherghi", "Svetlana Kiritchenko", "Swaroop Mishra", "Tal Linzen", "Tal Schuster", "Tao Li", "Tao Yu", "Tariq Ali", "Tatsu Hashimoto", "Te-Lin Wu", "Théo Desbordes", "Theodore Rothschild", "Thomas Phan", "Tianle Wang", "Tiberius Nkinyili", "Timo Schick", "Timofei Kornev", "Titus Tunduny", "Tobias Gerstenberg", "Trenton Chang", "Trishala Neeraj", "Tushar Khot", "Tyler Shultz", "Uri Shaham", "Vedant Misra", "Vera Demberg", "Victoria Nyamai", "Vikas Raunak", "Vinay Ramasesh", "Vinay Uday Prabhu", "Vishakh Padmakumar", "Vivek Srikumar", "William Fedus", "William Saunders", "William Zhang", "Wout Vossen", "Xiang Ren", "Xiaoyu Tong", "Xinran Zhao", "Xinyi Wu", "Xudong Shen", "Yadollah Yaghoobzadeh", "Yair Lakretz", "Yangqiu Song", "Yasaman Bahri", "Yejin Choi", "Yichi Yang", "Yiding Hao", "Yifu Chen", "Yonatan Belinkov", "Yu Hou", "Yufang Hou", "Yuntao Bai", "Zachary Seid", "Zhuoye Zhao", "Zijian Wang", "Zijie J. Wang", "Zirui Wang", "Ziyi Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "021e395904dacd8ca0479f037ff02b85f123af3fb44b271426f4074d5d4a03ec", "submitted_at": "2022-06-09T17:05:34Z", "updated_at": "2023-06-12T17:51:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2206.04615v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T08:01:28Z", "role": "primary", "sha256": "822cec31fa0db7ffa390fda806d88c2f9c0fe4d328742e7008b662916e92373a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2005.07648%2C2206.04615%2C2311.12022%2C2311.16922%2C2502.13130%2C2506.21506%2C2511.11332&max_results=7"}], "id": "arxiv:2206.04615", "identifiers": {"arxiv": "2206.04615"}, "kind": "paper", "last_verified_at": "2026-10-18T08:01:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2206.04615"}], "name": "Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-06-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-06-12", "url": "https://arxiv.org/abs/2206.04615"}
{"aliases": [], "authors": ["Andrea D'Ambrosio", "Miles Lifson", "Richard Linares"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2cabf479a39dea50343d2bc9d464a0659f426e555c68dc5316e86a468907b89f", "submitted_at": "2022-06-10T20:33:12Z", "updated_at": "2022-06-10T20:33:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2206.05345v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2206.05345", "identifiers": {"arxiv": "2206.05345"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2206.05345"}], "name": "The Capacity of Low Earth Orbit Computed using Source-sink Modeling", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-06-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-06-10", "url": "https://arxiv.org/abs/2206.05345"}
{"aliases": [], "authors": ["Chengyang Ying", "You Qiaoben", "Xinning Zhou", "Hang Su", "Wenbo Ding", "Jianyong Ai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf9009d1334588adf31def65453ae184b83353e2c9ad4fb78106a2b0198d5b17", "submitted_at": "2022-06-12T14:45:11Z", "updated_at": "2023-03-25T08:43:16Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2206.05751v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2206.05751", "identifiers": {"arxiv": "2206.05751"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2206.05751"}], "name": "Consistent Attack: Universal Adversarial Perturbation on Embodied Vision Navigation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-06-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-03-25", "url": "https://arxiv.org/abs/2206.05751"}
{"aliases": [], "authors": ["Linxi Fan", "Guanzhi Wang", "Yunfan Jiang", "Ajay Mandlekar", "Yuncong Yang", "Haoyi Zhu", "Andrew Tang", "De-An Huang", "Yuke Zhu", "Anima Anandkumar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f3008d72940cbc9819e77803bf1ae508ee6f112e16f6808e4a23111e210b81ab", "submitted_at": "2022-06-17T15:53:05Z", "updated_at": "2022-11-22T07:59:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2206.08853v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2206.08853", "identifiers": {"arxiv": "2206.08853"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2206.08853"}], "name": "MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledge", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-06-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-11-22", "url": "https://arxiv.org/abs/2206.08853"}
{"aliases": [], "authors": ["Disha Shrivastava", "Hugo Larochelle", "Daniel Tarlow"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2af01219f5119954d3568903179a5be96f2c856e5019043622bdf252909af6a", "submitted_at": "2022-06-26T10:51:25Z", "updated_at": "2023-06-05T18:43:50Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2206.12839v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2206.12839", "identifiers": {"arxiv": "2206.12839"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2206.12839"}], "name": "Repository-Level Prompt Generation for Large Language Models of Code", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-06-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-06-05", "url": "https://arxiv.org/abs/2206.12839"}
{"aliases": [], "authors": ["Andrey Malinin", "Andreas Athanasopoulos", "Muhamed Barakovic", "Meritxell Bach Cuadra", "Mark J. F. Gales", "Cristina Granziera", "Mara Graziani", "Nikolay Kartashev", "Konstantinos Kyriakopoulos", "Po-Jui Lu", "Nataliia Molchanova", "Antonis Nikitakis", "Vatsal Raina", "Francesco La Rosa", "Eli Sivena", "Vasileios Tsarsitalidis", "Efi Tsompopoulou", "Elena Volf"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "13be636542d44b01b8ffdfe4d7f73bc10727319b415a265df26d647f62d79d1d", "submitted_at": "2022-06-30T16:51:52Z", "updated_at": "2022-09-15T09:52:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2206.15407v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2206.15407", "identifiers": {"arxiv": "2206.15407"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2206.15407"}], "name": "Shifts 2.0: Extending The Dataset of Real Distributional Shifts", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-06-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-09-15", "url": "https://arxiv.org/abs/2206.15407"}
{"aliases": [], "authors": ["Shunyu Yao", "Howard Chen", "John Yang", "Karthik Narasimhan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6de4ab2a89875f1ccd99bbaa075c234d82979e845f27fe9530bdad0539c49570", "submitted_at": "2022-07-04T05:30:22Z", "updated_at": "2023-02-08T01:39:30Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2207.01206v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2207.01206", "identifiers": {"arxiv": "2207.01206"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2207.01206"}], "name": "WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2022-07-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Provides a simulated shopping environment where language agents search, configure, and select products from natural-language requirements.", "topics": ["evaluation", "web-navigation"], "updated_at": "2023-02-08", "url": "https://arxiv.org/abs/2207.01206"}
{"aliases": [], "authors": ["Zhipu Zhou", "Jie Chen", "Linwei Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d3b9d8120f170d9f14a43d260cdc02b38c1c67c27417d7e41567640f10fd07f", "submitted_at": "2022-07-11T22:20:51Z", "updated_at": "2022-07-11T22:20:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2207.05214v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2207.05214", "identifiers": {"arxiv": "2207.05214"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2207.05214"}], "name": "Shapley Computations Using Surrogate Model-Based Trees", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-07-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-07-11", "url": "https://arxiv.org/abs/2207.05214"}
{"aliases": [], "authors": ["Mukul Singh", "José Cambronero", "Sumit Gulwani", "Vu Le", "Carina Negreanu", "Mohammad Raza", "Gust Verbruggen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5afa77bf018aa741e03ccf27d11059657beaceba93d3a2172b57c0917b7304ee", "submitted_at": "2022-08-11T20:30:12Z", "updated_at": "2022-12-05T13:01:00Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2208.06032v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2208.06032", "identifiers": {"arxiv": "2208.06032"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2208.06032"}], "name": "CORNET: Learning Table Formatting Rules By Example", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-08-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-12-05", "url": "https://arxiv.org/abs/2208.06032"}
{"aliases": [], "authors": ["Prasanna Date", "Shruti Kulkarni", "Aaron Young", "Catherine Schuman", "Thomas Potok", "Jeffrey Vetter"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f73cd8b8905f4aafd18ca6a3cbd98c03b341727c2dd3e0fbb8d915741a1b2bf", "submitted_at": "2022-08-15T23:18:26Z", "updated_at": "2022-08-15T23:18:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2208.07468v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2208.07468", "identifiers": {"arxiv": "2208.07468"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2208.07468"}], "name": "Encoding Integers and Rationals on Neuromorphic Computers using Virtual Neuron", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-08-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-08-15", "url": "https://arxiv.org/abs/2208.07468"}
{"aliases": [], "authors": ["Dong Won Lee", "Chaitanya Ahuja", "Paul Pu Liang", "Sanika Natu", "Louis-Philippe Morency"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "516555992b8423e47be53e7e832cfff6dc5f40c1c2e04fe72acb20eee196c306", "submitted_at": "2022-08-17T05:30:18Z", "updated_at": "2022-08-17T05:30:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2208.08080v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2208.08080", "identifiers": {"arxiv": "2208.08080"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2208.08080"}], "name": "Multimodal Lecture Presentations Dataset: Understanding Multimodality in Educational Slides", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-08-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-08-17", "url": "https://arxiv.org/abs/2208.08080"}
{"aliases": [], "authors": ["Hudson Leone", "S Srikara", "Peter P. Rohde", "Simon Devitt"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aad3ddeba4cb300a117039b7680446651fee79257c5d4f2358373edfe6b12972", "submitted_at": "2022-08-31T23:14:45Z", "updated_at": "2024-02-15T05:57:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2209.00151v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2209.00151", "identifiers": {"arxiv": "2209.00151"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2209.00151"}], "name": "Upper Bounds for the Clock Speeds of Fault-Tolerant Distributed Quantum Computation using Satellites to Supply Entangled Photon Pairs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-08-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-15", "url": "https://arxiv.org/abs/2209.00151"}
{"aliases": [], "authors": ["Xingyuan Xu", "Weiwei Han", "Mengxi Tan", "Yang Sun", "Yang Li", "Jiayang Wu", "Roberto Morandotti", "Arnan Mitchell", "Kun Xu", "David J. Moss"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "18905f16d41664f97948524527c3cf3ee8b43f4c0c35add4515dd8e971b0c21e", "submitted_at": "2022-09-07T16:02:06Z", "updated_at": "2022-09-07T16:02:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2209.03252v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2209.03252", "identifiers": {"arxiv": "2209.03252"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2209.03252"}], "name": "Neuromorphic computing using wavelength-division multiplexing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-09-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-09-07", "url": "https://arxiv.org/abs/2209.03252"}
{"aliases": [], "authors": ["Yu-Chung Hsiao", "Fedir Zubach", "Gilles Baechler", "Srinivas Sunkara", "Victor Carbune", "Jason Lin", "Maria Wang", "Yun Zhu", "Jindong Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37f7ba5d5b9bbc96fd3a723e49190f296e08d2874dcd545272d3fcef3cc33d24", "submitted_at": "2022-09-16T23:49:00Z", "updated_at": "2025-02-09T21:09:17Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2209.08199v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2209.08199", "identifiers": {"arxiv": "2209.08199"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2209.08199"}], "name": "ScreenQA: Large-Scale Question-Answer Pairs over Mobile App Screenshots", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2022-09-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces screenshot question answering over mobile interfaces, linking answers to screen content and bounding boxes.", "topics": ["screen-understanding", "evaluation"], "updated_at": "2025-02-09", "url": "https://arxiv.org/abs/2209.08199"}
{"aliases": [], "authors": ["Bryan Wang", "Gang Li", "Yang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65813c1784ba00e354aff4e38796b1bdc8e9b691cba9dd56b8232b5d86555d1b", "submitted_at": "2022-09-18T20:58:39Z", "updated_at": "2023-02-17T21:24:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2209.08655v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2209.08655", "identifiers": {"arxiv": "2209.08655"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2209.08655"}], "name": "Enabling Conversational Interaction with Mobile UI using Large Language Models", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2022-09-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies few-shot prompting of a single language model for four mobile-UI interaction tasks, aiming to support conversational interface use without a separate trained model for each task.", "topics": ["mobile-navigation", "screen-understanding", "training-free"], "updated_at": "2023-02-17", "url": "https://arxiv.org/abs/2209.08655"}
{"aliases": [], "authors": ["Yuting Liu", "Albert Lee", "Kun Qian", "Peng Zhang", "Zhihua Xiao", "Haoran He", "Zheyu Ren", "Shun Kong Cheung", "Ruizi Liu", "Yaoyin Li", "Xu Zhang", "Zichao Ma", "Jianyuan Zhao", "Weiwei Zhao", "Guoqiang Yu", "Xin Wang", "Junwei Liu", "Zhongrui Wang", "Kang L. Wang", "Qiming Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47f45b2499c991923874b5a60893109193196e075ff7c4fa6c0425d74e89c186", "submitted_at": "2022-09-20T03:44:49Z", "updated_at": "2025-06-03T14:45:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2209.09443v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2209.09443", "identifiers": {"arxiv": "2209.09443"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2209.09443"}], "name": "Cryogenic in-memory computing using magnetic topological insulators", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-09-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-03", "url": "https://arxiv.org/abs/2209.09443"}
{"aliases": [], "authors": ["Akhil Francis", "Anjali A. Agrawal", "Jack H. Howard", "Efekan Kökcü", "A. F. Kemper"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86a34ca793d86d573d0246c40f84281aef399d4622107a9822d93781bf08de66", "submitted_at": "2022-09-21T18:01:33Z", "updated_at": "2022-09-21T18:01:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2209.10571v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2209.10571", "identifiers": {"arxiv": "2209.10571"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2209.10571"}], "name": "Subspace Diagonalization on Quantum Computers using Eigenvector Continuation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-09-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-09-21", "url": "https://arxiv.org/abs/2209.10571"}
{"aliases": [], "authors": ["Gang Li", "Yang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eaa351e34fa6bdcbe26aa6714eacda1f667411090abfdd5575735605dc89e864", "submitted_at": "2022-09-29T16:45:43Z", "updated_at": "2023-02-24T01:41:32Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2209.14927v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2209.14927", "identifiers": {"arxiv": "2209.14927"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2209.14927"}], "name": "Spotlight: Mobile UI Understanding using Vision-Language Models with a Focus", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2022-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Spotlight, a vision-language model conditioned on a UI screenshot and a selected region of interest. It evaluates screenshot-based mobile UI understanding and explores multitask learning and few-shot prompting.", "topics": ["screen-understanding", "grounding"], "updated_at": "2023-02-24", "url": "https://arxiv.org/abs/2209.14927"}
{"aliases": [], "authors": ["Jean Kaddour"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "859971891f6fb78d654a01547c254eeb231d1900aea6a7d85ca2d99cd7b0a9c3", "submitted_at": "2022-09-29T17:46:13Z", "updated_at": "2022-10-06T17:57:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2209.14981v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2209.14981", "identifiers": {"arxiv": "2209.14981"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2209.14981"}], "name": "Stop Wasting My Time! Saving Days of ImageNet and BERT Training with Latest Weight Averaging", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-10-06", "url": "https://arxiv.org/abs/2209.14981"}
{"aliases": [], "authors": ["Tao Li", "Gang Li", "Jingjie Zheng", "Purple Wang", "Yang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15e57e7dd30e6f3ae8b13cbd4a190b39339f2bbf7a551a6a5df5190dc62b5434", "submitted_at": "2022-09-29T21:08:18Z", "updated_at": "2022-09-29T21:08:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2209.15099v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2209.15099", "identifiers": {"arxiv": "2209.15099"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2209.15099"}], "name": "MUG: Interactive Multimodal Grounding on User Interfaces", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2022-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies iterative UI grounding in which users refine or correct an agent's interpretation over multiple conversational turns.", "topics": ["grounding", "human-in-the-loop"], "updated_at": "2022-09-29", "url": "https://arxiv.org/abs/2209.15099"}
{"aliases": [], "authors": ["Ray Chen Zheng", "Kaizhe Hu", "Zhecheng Yuan", "Boyuan Chen", "Huazhe Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "56ced20af8bc7eeb61c1fd5e373f63e5c536715066b3c81390aa1680f4c2de08", "submitted_at": "2022-10-04T04:42:26Z", "updated_at": "2023-03-01T08:01:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.01379v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2210.01379", "identifiers": {"arxiv": "2210.01379"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2210.01379"}], "name": "Extraneousness-Aware Imitation Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-10-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-03-01", "url": "https://arxiv.org/abs/2210.01379"}
{"aliases": [], "authors": ["Srinivas Sunkara", "Maria Wang", "Lijuan Liu", "Gilles Baechler", "Yu-Chung Hsiao", "Jindong", "Chen", "Abhanshu Sharma", "James Stout"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "24d037f1abeb7708852ac5517ce02cc46622e359d9feab407f188c0657cc72f4", "submitted_at": "2022-10-06T03:48:54Z", "updated_at": "2022-10-06T03:48:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2210.02663v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2210.02663", "identifiers": {"arxiv": "2210.02663"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2210.02663"}], "name": "Towards Better Semantic Understanding of Mobile Interfaces", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2022-10-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Augments RICO screens and view hierarchies with human labels for icon shape and meaning and for links between UI elements and text labels. It compares image-only and multimodal models on semantic understanding of previously unseen apps.", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2022-10-06", "url": "https://arxiv.org/abs/2210.02663"}
{"aliases": [], "authors": ["Kenton Lee", "Mandar Joshi", "Iulia Turc", "Hexiang Hu", "Fangyu Liu", "Julian Eisenschlos", "Urvashi Khandelwal", "Peter Shaw", "Ming-Wei Chang", "Kristina Toutanova"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7ebb243c4690aba5f1a92b0d1e050399560998a7f304ea483dae2f451cf51ec0", "submitted_at": "2022-10-07T06:42:06Z", "updated_at": "2023-06-15T21:34:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.03347v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2210.03347", "identifiers": {"arxiv": "2210.03347"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2210.03347"}], "name": "Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced. Adjacent visual-language pretraining: webpage screenshot parsing and explicit UI evaluation support interface understanding; the paper does not establish an autonomous GUI policy.", "published_at": "2022-10-07", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Pix2Struct pretrains an image-to-text model by reconstructing simplified HTML from masked webpage screenshots. Its variable-resolution visual representation is evaluated across several domains, including user-interface understanding.", "topics": ["screen-understanding", "data-generation"], "updated_at": "2023-06-15", "url": "https://arxiv.org/abs/2210.03347"}
{"aliases": [], "authors": ["Shunyu Yao", "Jeffrey Zhao", "Dian Yu", "Nan Du", "Izhak Shafran", "Karthik Narasimhan", "Yuan Cao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d54d4d76519ed21b2bfeedd94901a9bb559802e0c0bde7e09fb1e0a8bcdb3679", "submitted_at": "2022-10-06T01:00:32Z", "updated_at": "2023-03-10T01:00:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2210.03629v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2210.03629", "identifiers": {"arxiv": "2210.03629"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2210.03629"}], "name": "ReAct: Synergizing Reasoning and Acting in Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-10-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-03-10", "url": "https://arxiv.org/abs/2210.03629"}
{"aliases": [], "authors": ["Sunnie S. Y. Kim", "Elizabeth Anne Watkins", "Olga Russakovsky", "Ruth Fong", "Andrés Monroy-Hernández"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "125377dfdf1b0c59a3569cac9549c3813fe2ef986d8c387ed0e9acb6368d46a3", "submitted_at": "2022-10-02T20:17:11Z", "updated_at": "2023-02-16T20:14:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.03735v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2210.03735", "identifiers": {"arxiv": "2210.03735"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2210.03735"}], "name": "\"Help Me Help the AI\": Understanding How Explainability Can Support Human-AI Interaction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-10-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-02-16", "url": "https://arxiv.org/abs/2210.03735"}
{"aliases": [], "authors": ["Izzeddin Gur", "Ofir Nachum", "Yingjie Miao", "Mustafa Safdari", "Austin Huang", "Aakanksha Chowdhery", "Sharan Narang", "Noah Fiedel", "Aleksandra Faust"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b9089164b5c784735d8915596fe902840e01c8215f507caed0e10d4361d13ec", "submitted_at": "2022-10-08T07:27:17Z", "updated_at": "2023-05-19T17:42:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.03945v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2210.03945", "identifiers": {"arxiv": "2210.03945"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2210.03945"}], "name": "Understanding HTML with Large Language Models", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2022-10-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies language models for HTML element classification, page description and autonomous web navigation. Fine-tuned models are evaluated on MiniWoB, and a distilled, automatically labeled CommonCrawl HTML corpus supports the training approach.", "topics": ["screen-understanding", "web-navigation", "data-generation"], "updated_at": "2023-05-19", "url": "https://arxiv.org/abs/2210.03945"}
{"aliases": [], "authors": ["Kanishk Gandhi", "Siddharth Karamcheti", "Madeline Liao", "Dorsa Sadigh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b081a06d6d50afaf431c27dfb8ebcf8d28484a2cfd7c1401facc39914457a00", "submitted_at": "2022-10-14T19:37:55Z", "updated_at": "2022-10-14T19:37:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2210.08073v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2210.08073", "identifiers": {"arxiv": "2210.08073"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2210.08073"}], "name": "Eliciting Compatible Demonstrations for Multi-Human Imitation Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-10-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-10-14", "url": "https://arxiv.org/abs/2210.08073"}
{"aliases": [], "authors": ["Luis A. Martínez-Martínez", "Tzu-Ching Yen", "Artur F. Izmaylov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc77763eae9c5c80f7dc12f806ed73f708240f4f64d8d59c615eb620c0101c53", "submitted_at": "2022-10-18T22:22:54Z", "updated_at": "2023-08-07T22:32:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.10189v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2210.10189", "identifiers": {"arxiv": "2210.10189"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2210.10189"}], "name": "Assessment of various Hamiltonian partitionings for the electronic structure problem on a quantum computer using the Trotter approximation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-10-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-07", "url": "https://arxiv.org/abs/2210.10189"}
{"aliases": [], "authors": ["Sean Kulinski", "David I. Inouye"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7a3b5c10e9a88cbbc4307d137c24eb66a998d013a1880e2773e62e969cfd6abb", "submitted_at": "2022-10-19T03:38:57Z", "updated_at": "2023-06-20T04:30:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.10275v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2210.10275", "identifiers": {"arxiv": "2210.10275"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2210.10275"}], "name": "Towards Explaining Distribution Shifts", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-06-20", "url": "https://arxiv.org/abs/2210.10275"}
{"aliases": [], "authors": ["Soroush Nasiriany", "Tian Gao", "Ajay Mandlekar", "Yuke Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb18a010cf6e2219314db4b68c24c4789f1f84d13e011b799ac89c95c53beaa2", "submitted_at": "2022-10-20T17:34:59Z", "updated_at": "2022-11-15T00:13:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.11435v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2210.11435", "identifiers": {"arxiv": "2210.11435"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2210.11435"}], "name": "Learning and Retrieval from Prior Data for Skill-based Imitation Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-11-15", "url": "https://arxiv.org/abs/2210.11435"}
{"aliases": [], "authors": ["Juha Leppäkangas", "Nicolas Vogt", "Keith R. Fratus", "Kirsten Bark", "Jesse A. Vaitkus", "Pascal Stadler", "Jan-Michael Reiner", "Sebastian Zanker", "Michael Marthaler"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0dc16e6d55eebd54b326ffe9bbdea69d7fd050a05c2f56151f500b5ef1d8ab7a", "submitted_at": "2022-10-21T17:47:32Z", "updated_at": "2023-12-11T14:39:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2210.12138v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2210.12138", "identifiers": {"arxiv": "2210.12138"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2210.12138"}], "name": "A quantum algorithm for solving open system dynamics on quantum computers using noise", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-10-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-11", "url": "https://arxiv.org/abs/2210.12138"}
{"aliases": [], "authors": ["Sunil Pai", "Olav Solgaard", "Shanhui Fan", "David A. B. Miller"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "edc1a293c0b826f47ae5ede3e7c9b9d6e0c66d6b323de6d3bc3474850b0ac404", "submitted_at": "2022-10-30T19:41:33Z", "updated_at": "2022-10-30T19:41:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2210.16935v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2210.16935", "identifiers": {"arxiv": "2210.16935"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2210.16935"}], "name": "Scalable and self-correcting photonic computation using balanced photonic binary tree cascades", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-10-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-10-30", "url": "https://arxiv.org/abs/2210.16935"}
{"aliases": [], "authors": ["Sagar Gubbi Venkatesh", "Partha Talukdar", "Srini Narayanan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52319819402cec76a69c0bb513f3a7000053208ddd9d2a8c675e055422402fbe", "submitted_at": "2022-11-14T18:36:19Z", "updated_at": "2023-05-23T16:08:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2211.07615v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2211.07615", "identifiers": {"arxiv": "2211.07615"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2211.07615"}], "name": "UGIF: UI Grounded Instruction Following", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2022-11-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces multilingual smartphone instruction-following data and a pipeline that retrieves help instructions and maps them to UI actions.", "topics": ["mobile-navigation", "grounding"], "updated_at": "2023-05-23", "url": "https://arxiv.org/abs/2211.07615"}
{"aliases": [], "authors": ["Chih-Hui Ho", "Srikar Appalaraju", "Bhavan Jasani", "R. Manmatha", "Nuno Vasconcelos"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6198cd83315c01e6269eb16ffbd017c85eca84c0da337bb5d50665e065c66a2c", "submitted_at": "2022-11-15T05:34:40Z", "updated_at": "2022-11-15T05:34:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2211.07912v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2211.07912", "identifiers": {"arxiv": "2211.07912"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2211.07912"}], "name": "YORO -- Lightweight End to End Visual Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-11-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-11-15", "url": "https://arxiv.org/abs/2211.07912"}
{"aliases": [], "authors": ["Niklas Budinger", "Akira Furusawa", "Peter van Loock"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c1808982568795d288e32bb3911cb685b6bd55a123df86d08c146f19c6a3c6ae", "submitted_at": "2022-11-16T17:21:30Z", "updated_at": "2024-07-13T21:17:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2211.09060v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2211.09060", "identifiers": {"arxiv": "2211.09060"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2211.09060"}], "name": "All-optical quantum computing using cubic phase gates", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-11-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-13", "url": "https://arxiv.org/abs/2211.09060"}
{"aliases": [], "authors": ["Fábio Perez", "Ian Ribeiro"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "639f18c9229434c07a0455ccee3410f09738d2c88f1c21395f98bf484193cdc8", "submitted_at": "2022-11-17T13:43:20Z", "updated_at": "2022-11-17T13:43:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2211.09527v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2211.09527", "identifiers": {"arxiv": "2211.09527"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2211.09527"}], "name": "Ignore Previous Prompt: Attack Techniques For Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-11-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-11-17", "url": "https://arxiv.org/abs/2211.09527"}
{"aliases": [], "authors": ["Joseph D. Hart", "Francesco Sorrentino", "Thomas L. Carroll"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "02c684d2a0f41a902b37c4174aebdc72ef6b002bb5af4f0327edb32fa1f1780b", "submitted_at": "2022-11-29T14:55:51Z", "updated_at": "2023-04-25T22:24:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2211.17095v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2211.17095", "identifiers": {"arxiv": "2211.17095"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2211.17095"}], "name": "Time-shift selection for reservoir computing using a rank-revealing QR algorithm", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-11-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-04-25", "url": "https://arxiv.org/abs/2211.17095"}
{"aliases": [], "authors": ["Shreshth Tuli", "Giuliano Casale", "Ludmila Cherkasova", "Nicholas R. Jennings"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51fef00bdd85f0c17fade5a4f567f178b6fb7a994fe36bc38912d2eb34af2861", "submitted_at": "2022-12-02T16:51:58Z", "updated_at": "2022-12-02T16:51:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2212.01302v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2212.01302", "identifiers": {"arxiv": "2212.01302"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2212.01302"}], "name": "DeepFT: Fault-Tolerant Edge Computing using a Self-Supervised Deep Surrogate Model", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-12-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-12-02", "url": "https://arxiv.org/abs/2212.01302"}
{"aliases": [], "authors": ["Luis Maßny", "Antonia Wachter-Zeh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e9e9c926dfe176758a299ff28a578c054cef1fb48ee99e4a07bf61a0f8b6f20", "submitted_at": "2022-12-08T14:30:59Z", "updated_at": "2022-12-22T16:13:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2212.04288v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2212.04288", "identifiers": {"arxiv": "2212.04288"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2212.04288"}], "name": "Secure Over-the-Air Computation using Zero-Forced Artificial Noise", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-12-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-12-22", "url": "https://arxiv.org/abs/2212.04288"}
{"aliases": [], "authors": ["Nils Quetschlich", "Lukas Burgholzer", "Robert Wille"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f76992f582117cb3659bf12bfa610744588c04b3e241b9b1ed104da069568c53", "submitted_at": "2022-12-08T19:00:01Z", "updated_at": "2023-04-04T09:28:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2212.04508v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2212.04508", "identifiers": {"arxiv": "2212.04508"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2212.04508"}], "name": "Compiler Optimization for Quantum Computing Using Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-12-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-04-04", "url": "https://arxiv.org/abs/2212.04508"}
{"aliases": [], "authors": ["N. Rasha Shanaz", "K. Murali", "P. Muruganandam"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f7dc71c9a60ce1f2ed7e0a13e0f7809a53724e89bb318b34e37b19dc89d68bde", "submitted_at": "2022-12-17T00:25:56Z", "updated_at": "2022-12-17T00:25:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2212.11141v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2212.11141", "identifiers": {"arxiv": "2212.11141"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2212.11141"}], "name": "Reservoir Computing Using Complex Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-12-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2022-12-17", "url": "https://arxiv.org/abs/2212.11141"}
{"aliases": [], "authors": ["Federico Malato", "Florian Leopold", "Amogh Raut", "Ville Hautamäki", "Andrew Melnik"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6ce03870cd9bf085cd9b8c13e907d11ebdaa910238abdd411486f7c71de00b5", "submitted_at": "2022-12-27T00:20:37Z", "updated_at": "2023-04-17T05:38:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2212.13326v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2212.13326", "identifiers": {"arxiv": "2212.13326"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2212.13326"}], "name": "Behavioral Cloning via Search in Video PreTraining Latent Space", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-12-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-04-17", "url": "https://arxiv.org/abs/2212.13326"}
{"aliases": [], "authors": ["Dandison Ukpabi", "Heikki Karjaluoto", "Astrid Bötticher", "Anastasija Nikiforova", "Dragoş Petrescu", "Paulina Schindler", "Visvaldis Valtenbergs", "Lennard Lehmann", "Abuzer Yakaryilmaz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b365a86895c4b9e6e3c45f41b2e02847f0ca90659ba833d99ec49c4e5c01a4d2", "submitted_at": "2022-12-19T12:27:12Z", "updated_at": "2023-01-05T11:35:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2212.13909v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2212.13909", "identifiers": {"arxiv": "2212.13909"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2212.13909"}], "name": "Framework for understanding quantum computing use cases from a multidisciplinary perspective and future research directions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2022-12-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-01-05", "url": "https://arxiv.org/abs/2212.13909"}
{"aliases": [], "authors": ["M. Mazzanti", "R. Gerritsma", "R. J. C. Spreeuw", "A. Safavi-Naini"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a4c9544ad60f0958a9d701de430c0c0d070bb8d90c1d23e01bcd8b92ba3e42b4", "submitted_at": "2023-01-11T19:00:09Z", "updated_at": "2023-01-11T19:00:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2301.04668v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2301.04668", "identifiers": {"arxiv": "2301.04668"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2301.04668"}], "name": "Trapped Ion Quantum Computing using Optical Tweezers and the Magnus Effect", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-01-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-01-11", "url": "https://arxiv.org/abs/2301.04668"}
{"aliases": [], "authors": ["Frank Zhigang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c352a47f6dd92d9900daf0452bac35e595e3d1fba8784b1c312bef3bea689528", "submitted_at": "2023-01-18T18:24:16Z", "updated_at": "2023-06-03T05:55:16Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2301.09418v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2301.09418", "identifiers": {"arxiv": "2301.09418"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2301.09418"}], "name": "Near-Landauer-Bound Quantum Computing Using Single Spins", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-01-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-06-03", "url": "https://arxiv.org/abs/2301.09418"}
{"aliases": [], "authors": ["Pratyay Banerjee", "Shweti Mahajan", "Kushal Arora", "Chitta Baral", "Oriana Riva"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb647a96c7f8263786e85a5061c8c5284cb92ed0b2ebc4c405e1243b2ca3fa2b", "submitted_at": "2023-01-23T09:05:49Z", "updated_at": "2023-01-23T09:05:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2301.10165v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2301.10165", "identifiers": {"arxiv": "2301.10165"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2301.10165"}], "name": "Lexi: Self-Supervised Learning of the UI Language", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2023-01-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Lexi and UICaption to learn UI image-text representations from descriptions of interface functionality. Evaluation covers UI action entailment, instruction-based retrieval, referring-expression grounding, and UI entity recognition.", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2023-01-23", "url": "https://arxiv.org/abs/2301.10165"}
{"aliases": [], "authors": ["Tim Pearce", "Tabish Rashid", "Anssi Kanervisto", "Dave Bignell", "Mingfei Sun", "Raluca Georgescu", "Sergio Valcarcel Macua", "Shan Zheng Tan", "Ida Momennejad", "Katja Hofmann", "Sam Devlin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f3fe6fc5e04810b8332605001b6da24b76d06d224f6b106e97561772ff55028", "submitted_at": "2023-01-25T16:31:05Z", "updated_at": "2023-03-03T14:18:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2301.10677v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2301.10677", "identifiers": {"arxiv": "2301.10677"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2301.10677"}], "name": "Imitating Human Behaviour with Diffusion Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-01-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-03-03", "url": "https://arxiv.org/abs/2301.10677"}
{"aliases": [], "authors": ["Jason Wu", "Siyan Wang", "Siman Shen", "Yi-Hao Peng", "Jeffrey Nichols", "Jeffrey P. Bigham"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0cc3b91aa85bc8c7d88a878c7b24926943eea0bcc92e25f517915242f2a82a39", "submitted_at": "2023-01-30T20:47:12Z", "updated_at": "2023-01-30T20:47:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2301.13280v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2301.13280", "identifiers": {"arxiv": "2301.13280"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2301.13280"}], "name": "WebUI: A Dataset for Enhancing Visual UI Understanding with Web Semantics", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-01-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Collects rendered webpages and structural metadata for visual UI modeling, studying how web semantics transfer to interface-understanding tasks.", "topics": ["screen-understanding", "data-generation"], "updated_at": "2023-01-30", "url": "https://arxiv.org/abs/2301.13280"}
{"aliases": [], "authors": ["Harshit Joshi", "Abishai Ebenezer", "José Cambronero", "Sumit Gulwani", "Aditya Kanade", "Vu Le", "Ivan Radiček", "Gust Verbruggen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff6e0dbc7f7ba22430915cac58358a3afe5c6167beecfca06acdf3f8b0f15859", "submitted_at": "2023-01-31T17:29:43Z", "updated_at": "2023-12-19T22:56:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2301.13779v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2301.13779", "identifiers": {"arxiv": "2301.13779"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2301.13779"}], "name": "FLAME: A small language model for spreadsheet formulas", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-01-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-19", "url": "https://arxiv.org/abs/2301.13779"}
{"aliases": [], "authors": ["Zihao Wang", "Shaofei Cai", "Guanzhou Chen", "Anji Liu", "Xiaojian Ma", "Yitao Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "85475d5b76011627e8d576f316710b84cc2819cd1d686b322e1939402fa24b42", "submitted_at": "2023-02-03T06:06:27Z", "updated_at": "2024-07-08T05:56:47Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2302.01560v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2302.01560", "identifiers": {"arxiv": "2302.01560"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2302.01560"}], "name": "Describe, Explain, Plan and Select: Interactive Planning with Large Language Models Enables Open-World Multi-Task Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-02-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-08", "url": "https://arxiv.org/abs/2302.01560"}
{"aliases": [], "authors": ["Yuxin Wen", "Neel Jain", "John Kirchenbauer", "Micah Goldblum", "Jonas Geiping", "Tom Goldstein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "10b5f95064b8c793c7c18c2e9a1580465e46817ddbb1e4f178e12f28a2575c3a", "submitted_at": "2023-02-07T18:40:18Z", "updated_at": "2023-06-01T12:26:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2302.03668v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2302.03668", "identifiers": {"arxiv": "2302.03668"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2302.03668"}], "name": "Hard Prompts Made Easy: Gradient-Based Discrete Optimization for Prompt Tuning and Discovery", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-02-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-06-01", "url": "https://arxiv.org/abs/2302.03668"}
{"aliases": [], "authors": ["Terufumi Yamaguchi", "Sumito Tsunegi", "Kohei Nakajima", "Tomohiro Taniguchi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d9eb32a8c5d427c162222dfc93360c6258573fb870fb74956bcd90e4426ba1eb", "submitted_at": "2023-02-07T22:00:18Z", "updated_at": "2023-02-07T22:00:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2302.03769v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2302.03769", "identifiers": {"arxiv": "2302.03769"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2302.03769"}], "name": "Computational capability for physical reservoir computing using a spin-torque oscillator with two free layers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-02-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-02-07", "url": "https://arxiv.org/abs/2302.03769"}
{"aliases": [], "authors": ["Timo Schick", "Jane Dwivedi-Yu", "Roberto Dessì", "Roberta Raileanu", "Maria Lomeli", "Luke Zettlemoyer", "Nicola Cancedda", "Thomas Scialom"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d131ed654bcc879a2cc35030ad2658788743592dd5f0e10697fb2de943c4014e", "submitted_at": "2023-02-09T16:49:57Z", "updated_at": "2023-02-09T16:49:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2302.04761v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2302.04761", "identifiers": {"arxiv": "2302.04761"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2302.04761"}], "name": "Toolformer: Language Models Can Teach Themselves to Use Tools", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-02-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-02-09", "url": "https://arxiv.org/abs/2302.04761"}
{"aliases": [], "authors": ["Xiaoguo Li", "Bowen Zhao", "Guomin Yang", "Tao Xiang", "Jian Weng", "Robert H. Deng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "84d589792ed9edffdec9ca1a81b044d222463e824a64593ef3671009610fe598", "submitted_at": "2023-02-23T16:33:56Z", "updated_at": "2023-02-23T16:33:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2302.12150v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2302.12150", "identifiers": {"arxiv": "2302.12150"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2302.12150"}], "name": "A Survey of Secure Computation Using Trusted Execution Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-02-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-02-23", "url": "https://arxiv.org/abs/2302.12150"}
{"aliases": [], "authors": ["Kai Greshake", "Sahar Abdelnabi", "Shailesh Mishra", "Christoph Endres", "Thorsten Holz", "Mario Fritz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ca4788ef05f4ccdb485c253487ab1db44faf9dc9e1f8c42360ad28f674df10f9", "submitted_at": "2023-02-23T17:14:38Z", "updated_at": "2023-05-05T14:26:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2302.12173v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2302.12173", "identifiers": {"arxiv": "2302.12173"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2302.12173"}], "name": "Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-02-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-05", "url": "https://arxiv.org/abs/2302.12173"}
{"aliases": [], "authors": ["Minae Kwon", "Sang Michael Xie", "Kalesha Bullard", "Dorsa Sadigh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b5cde34c777ee3b16cba1a478a6be5e1927993022fcd3b7ad843b6c6cbd030cc", "submitted_at": "2023-02-27T22:09:35Z", "updated_at": "2023-02-27T22:09:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2303.00001v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2303.00001", "identifiers": {"arxiv": "2303.00001"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.00001"}], "name": "Reward Design with Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-02-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-02-27", "url": "https://arxiv.org/abs/2303.00001"}
{"aliases": [], "authors": ["Oscar Lee", "Robin Msiska", "Maarten A. Brems", "Mathias Klaui", "Hidekazu Kurebayashi", "Karin Everschor-Sitte"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77f649ac9e825553091236c5c0c04f26f033c8c02627e3484d36e6dc06649280", "submitted_at": "2023-03-01T18:06:26Z", "updated_at": "2023-05-21T10:44:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2303.00708v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2303.00708", "identifiers": {"arxiv": "2303.00708"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.00708"}], "name": "Perspective on unconventional computing using magnetic skyrmions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-21", "url": "https://arxiv.org/abs/2303.00708"}
{"aliases": [], "authors": ["Erika Kawakami", "Jiabao Chen", "Mónica Benito", "Denis Konstantinov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c6c35a50d40ec61800596081f99b24b7bf8402293f799886f505b71686b99405", "submitted_at": "2023-03-07T07:13:04Z", "updated_at": "2023-10-19T04:31:35Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2303.03688v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2303.03688", "identifiers": {"arxiv": "2303.03688"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.03688"}], "name": "Blueprint for quantum computing using electrons on helium", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-19", "url": "https://arxiv.org/abs/2303.03688"}
{"aliases": [], "authors": ["Venkatesh Vadde", "Bhaskaran Muralidharan", "Abhishek Sharma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7feb19db8c64b32d373a12d53c79a6b3c1df098eec6c7cc06ce61376ff1e08e3", "submitted_at": "2023-03-11T17:27:22Z", "updated_at": "2023-03-11T17:27:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2303.06463v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2303.06463", "identifiers": {"arxiv": "2303.06463"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.06463"}], "name": "Power efficient ReLU design for neuromorphic computing using spin Hall effect", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-03-11", "url": "https://arxiv.org/abs/2303.06463"}
{"aliases": [], "authors": ["Yuqing Du", "Ksenia Konyushkova", "Misha Denil", "Akhil Raju", "Jessica Landon", "Felix Hill", "Nando de Freitas", "Serkan Cabi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "475860ad80e41d2cfc2f48a22c97d19f678445b1eb07b68bd0f66cce8647b3ce", "submitted_at": "2023-03-13T16:54:11Z", "updated_at": "2023-03-13T16:54:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2303.07280v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2303.07280", "identifiers": {"arxiv": "2303.07280"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.07280"}], "name": "Vision-Language Models as Success Detectors", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-03-13", "url": "https://arxiv.org/abs/2303.07280"}
{"aliases": [], "authors": ["OpenAI", "Josh Achiam", "Steven Adler", "Sandhini Agarwal", "Lama Ahmad", "Ilge Akkaya", "Florencia Leoni Aleman", "Diogo Almeida", "Janko Altenschmidt", "Sam Altman", "Shyamal Anadkat", "Red Avila", "Igor Babuschkin", "Suchir Balaji", "Valerie Balcom", "Paul Baltescu", "Haiming Bao", "Mohammad Bavarian", "Jeff Belgum", "Irwan Bello", "Jake Berdine", "Gabriel Bernadett-Shapiro", "Christopher Berner", "Lenny Bogdonoff", "Oleg Boiko", "Madelaine Boyd", "Anna-Luisa Brakman", "Greg Brockman", "Tim Brooks", "Miles Brundage", "Kevin Button", "Trevor Cai", "Rosie Campbell", "Andrew Cann", "Brittany Carey", "Chelsea Carlson", "Rory Carmichael", "Brooke Chan", "Che Chang", "Fotis Chantzis", "Derek Chen", "Sully Chen", "Ruby Chen", "Jason Chen", "Mark Chen", "Ben Chess", "Chester Cho", "Casey Chu", "Hyung Won Chung", "Dave Cummings", "Jeremiah Currier", "Yunxing Dai", "Cory Decareaux", "Thomas Degry", "Noah Deutsch", "Damien Deville", "Arka Dhar", "David Dohan", "Steve Dowling", "Sheila Dunning", "Adrien Ecoffet", "Atty Eleti", "Tyna Eloundou", "David Farhi", "Liam Fedus", "Niko Felix", "Simón Posada Fishman", "Juston Forte", "Isabella Fulford", "Leo Gao", "Elie Georges", "Christian Gibson", "Vik Goel", "Tarun Gogineni", "Gabriel Goh", "Rapha Gontijo-Lopes", "Jonathan Gordon", "Morgan Grafstein", "Scott Gray", "Ryan Greene", "Joshua Gross", "Shixiang Shane Gu", "Yufei Guo", "Chris Hallacy", "Jesse Han", "Jeff Harris", "Yuchen He", "Mike Heaton", "Johannes Heidecke", "Chris Hesse", "Alan Hickey", "Wade Hickey", "Peter Hoeschele", "Brandon Houghton", "Kenny Hsu", "Shengli Hu", "Xin Hu", "Joost Huizinga", "Shantanu Jain", "Shawn Jain", "Joanne Jang", "Angela Jiang", "Roger Jiang", "Haozhun Jin", "Denny Jin", "Shino Jomoto", "Billie Jonn", "Heewoo Jun", "Tomer Kaftan", "Łukasz Kaiser", "Ali Kamali", "Ingmar Kanitscheider", "Nitish Shirish Keskar", "Tabarak Khan", "Logan Kilpatrick", "Jong Wook Kim", "Christina Kim", "Yongjik Kim", "Jan Hendrik Kirchner", "Jamie Kiros", "Matt Knight", "Daniel Kokotajlo", "Łukasz Kondraciuk", "Andrew Kondrich", "Aris Konstantinidis", "Kyle Kosic", "Gretchen Krueger", "Vishal Kuo", "Michael Lampe", "Ikai Lan", "Teddy Lee", "Jan Leike", "Jade Leung", "Daniel Levy", "Chak Ming Li", "Rachel Lim", "Molly Lin", "Stephanie Lin", "Mateusz Litwin", "Theresa Lopez", "Ryan Lowe", "Patricia Lue", "Anna Makanju", "Kim Malfacini", "Sam Manning", "Todor Markov", "Yaniv Markovski", "Bianca Martin", "Katie Mayer", "Andrew Mayne", "Bob McGrew", "Scott Mayer McKinney", "Christine McLeavey", "Paul McMillan", "Jake McNeil", "David Medina", "Aalok Mehta", "Jacob Menick", "Luke Metz", "Andrey Mishchenko", "Pamela Mishkin", "Vinnie Monaco", "Evan Morikawa", "Daniel Mossing", "Tong Mu", "Mira Murati", "Oleg Murk", "David Mély", "Ashvin Nair", "Reiichiro Nakano", "Rajeev Nayak", "Arvind Neelakantan", "Richard Ngo", "Hyeonwoo Noh", "Long Ouyang", "Cullen O'Keefe", "Jakub Pachocki", "Alex Paino", "Joe Palermo", "Ashley Pantuliano", "Giambattista Parascandolo", "Joel Parish", "Emy Parparita", "Alex Passos", "Mikhail Pavlov", "Andrew Peng", "Adam Perelman", "Filipe de Avila Belbute Peres", "Michael Petrov", "Henrique Ponde de Oliveira Pinto", "Michael", "Pokorny", "Michelle Pokrass", "Vitchyr H. Pong", "Tolly Powell", "Alethea Power", "Boris Power", "Elizabeth Proehl", "Raul Puri", "Alec Radford", "Jack Rae", "Aditya Ramesh", "Cameron Raymond", "Francis Real", "Kendra Rimbach", "Carl Ross", "Bob Rotsted", "Henri Roussez", "Nick Ryder", "Mario Saltarelli", "Ted Sanders", "Shibani Santurkar", "Girish Sastry", "Heather Schmidt", "David Schnurr", "John Schulman", "Daniel Selsam", "Kyla Sheppard", "Toki Sherbakov", "Jessica Shieh", "Sarah Shoker", "Pranav Shyam", "Szymon Sidor", "Eric Sigler", "Maddie Simens", "Jordan Sitkin", "Katarina Slama", "Ian Sohl", "Benjamin Sokolowsky", "Yang Song", "Natalie Staudacher", "Felipe Petroski Such", "Natalie Summers", "Ilya Sutskever", "Jie Tang", "Nikolas Tezak", "Madeleine B. Thompson", "Phil Tillet", "Amin Tootoonchian", "Elizabeth Tseng", "Preston Tuggle", "Nick Turley", "Jerry Tworek", "Juan Felipe Cerón Uribe", "Andrea Vallone", "Arun Vijayvergiya", "Chelsea Voss", "Carroll Wainwright", "Justin Jay Wang", "Alvin Wang", "Ben Wang", "Jonathan Ward", "Jason Wei", "CJ Weinmann", "Akila Welihinda", "Peter Welinder", "Jiayi Weng", "Lilian Weng", "Matt Wiethoff", "Dave Willner", "Clemens Winter", "Samuel Wolrich", "Hannah Wong", "Lauren Workman", "Sherwin Wu", "Jeff Wu", "Michael Wu", "Kai Xiao", "Tao Xu", "Sarah Yoo", "Kevin Yu", "Qiming Yuan", "Wojciech Zaremba", "Rowan Zellers", "Chong Zhang", "Marvin Zhang", "Shengjia Zhao", "Tianhao Zheng", "Juntang Zhuang", "William Zhuk", "Barret Zoph"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7845a28b1cb12e4366e97338f919ba33bdb2f2fd1a45db6990bee4faec62832e", "submitted_at": "2023-03-15T17:15:04Z", "updated_at": "2024-03-04T06:01:33Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2303.08774v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2303.08774", "identifiers": {"arxiv": "2303.08774"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.08774"}], "name": "GPT-4 Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-04", "url": "https://arxiv.org/abs/2303.08774"}
{"aliases": [], "authors": ["Noah Shinn", "Federico Cassano", "Edward Berman", "Ashwin Gopinath", "Karthik Narasimhan", "Shunyu Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "316e3d252ba03a563d838cbb67f8b46324fa4ce37ebafa4573b6efb0f16155d6", "submitted_at": "2023-03-20T18:08:50Z", "updated_at": "2023-10-10T05:21:45Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2303.11366v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2303.11366", "identifiers": {"arxiv": "2303.11366"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.11366"}], "name": "Reflexion: Language Agents with Verbal Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-10", "url": "https://arxiv.org/abs/2303.11366"}
{"aliases": [], "authors": ["Li Zhou", "Zikun Zhou", "Kaige Mao", "Zhenyu He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "31370e2b0963c4db283a94d46de557cee285b5a1280663deefa131695c673d3f", "submitted_at": "2023-03-21T17:09:03Z", "updated_at": "2023-03-21T17:09:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2303.12027v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2303.12027", "identifiers": {"arxiv": "2303.12027"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.12027"}], "name": "Joint Visual Grounding and Tracking with Natural Language Specification", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-03-21", "url": "https://arxiv.org/abs/2303.12027"}
{"aliases": [], "authors": ["Haoqi Yuan", "Chi Zhang", "Hongcheng Wang", "Feiyang Xie", "Penglin Cai", "Hao Dong", "Zongqing Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bc98dfc5ab62a011837395e37de215b6cfba653c3d022296dd74a5b6bade1ff3", "submitted_at": "2023-03-29T09:45:50Z", "updated_at": "2023-12-04T14:53:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2303.16563v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2303.16563", "identifiers": {"arxiv": "2303.16563"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.16563"}], "name": "Skill Reinforcement Learning and Planning for Open-World Long-Horizon Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-04", "url": "https://arxiv.org/abs/2303.16563"}
{"aliases": [], "authors": ["Zequn Cao", "Xiaoheng Deng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8c5ec98167dd51f96e66cee1dff03713fd2a4d279da0ccff7b34659646886acc", "submitted_at": "2023-03-30T02:18:51Z", "updated_at": "2023-03-30T02:18:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2303.17100v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2303.17100", "identifiers": {"arxiv": "2303.17100"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.17100"}], "name": "Dependent Task Offloading in Edge Computing Using GNN and Deep Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-03-30", "url": "https://arxiv.org/abs/2303.17100"}
{"aliases": [], "authors": ["Geunwoo Kim", "Pierre Baldi", "Stephen McAleer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8f36bd6088175b337b750d91d0b7c04d2c156742e5c4090a74f00a482157e386", "submitted_at": "2023-03-30T16:01:52Z", "updated_at": "2023-11-16T20:15:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2303.17491v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2303.17491", "identifiers": {"arxiv": "2303.17491"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2303.17491"}], "name": "Language Models can Solve Computer Tasks", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2023-03-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces recursive criticism and improvement prompting for language-model agents performing computer tasks, evaluated on MiniWoB++ with a small number of demonstrations per task.", "topics": ["web-navigation", "self-reflection", "training-free"], "updated_at": "2023-11-16", "url": "https://arxiv.org/abs/2303.17491"}
{"aliases": [], "authors": ["Yongliang Shen", "Kaitao Song", "Xu Tan", "Dongsheng Li", "Weiming Lu", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ddab3ad61a046fe11ddf68025292f3d11b59ce43662479c06f5075267f04414", "submitted_at": "2023-03-30T17:48:28Z", "updated_at": "2023-12-03T18:17:21Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2303.17580v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2303.17580", "identifiers": {"arxiv": "2303.17580"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.17580"}], "name": "HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-03", "url": "https://arxiv.org/abs/2303.17580"}
{"aliases": [], "authors": ["Guohao Li", "Hasan Abed Al Kader Hammoud", "Hani Itani", "Dmitrii Khizbullin", "Bernard Ghanem"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6cc05737de8b8876ba8abef085b0ffb777377971da7f2f3934152351c7c0d674", "submitted_at": "2023-03-31T01:09:00Z", "updated_at": "2023-11-02T17:34:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2303.17760v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2303.17760", "identifiers": {"arxiv": "2303.17760"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2303.17760"}], "name": "CAMEL: Communicative Agents for \"Mind\" Exploration of Large Language Model Society", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-03-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-02", "url": "https://arxiv.org/abs/2303.17760"}
{"aliases": [], "authors": ["Joon Sung Park", "Joseph C. O'Brien", "Carrie J. Cai", "Meredith Ringel Morris", "Percy Liang", "Michael S. Bernstein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9205ed79d058490ec68d28a5c70fff7f8ca61e4e3d6d01a39349e917c9d973c", "submitted_at": "2023-04-07T01:55:19Z", "updated_at": "2023-08-06T00:21:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2304.03442v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}], "id": "arxiv:2304.03442", "identifiers": {"arxiv": "2304.03442"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2304.03442"}], "name": "Generative Agents: Interactive Simulacra of Human Behavior", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-04-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-06", "url": "https://arxiv.org/abs/2304.03442"}
{"aliases": [], "authors": ["Yingqiang Ge", "Wenyue Hua", "Kai Mei", "Jianchao Ji", "Juntao Tan", "Shuyuan Xu", "Zelong Li", "Yongfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5197ad473544cfb8b6565408fbd467db4d0c1e1ede313e6308ec65ccfcac469", "submitted_at": "2023-04-10T03:55:35Z", "updated_at": "2023-11-03T15:24:18Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2304.04370v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2304.04370", "identifiers": {"arxiv": "2304.04370"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2304.04370"}], "name": "OpenAGI: When LLM Meets Domain Experts", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-04-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-03", "url": "https://arxiv.org/abs/2304.04370"}
{"aliases": [], "authors": ["Hao Wen", "Hongming Wang", "Jiaxuan Liu", "Yuanchun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "420c9005fd0c687c27c758da70ef05a60b3b2d9b9efbf91a5d7661bf893a8b09", "submitted_at": "2023-04-14T11:31:56Z", "updated_at": "2024-01-07T08:23:47Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2304.07061v5"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2304.07061", "identifiers": {"arxiv": "2304.07061"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2304.07061"}], "name": "DroidBot-GPT: GPT-powered UI Automation for Android", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-04-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Transforms Android UI states and available actions into language prompts for an LLM to choose and execute task steps.", "topics": ["mobile-navigation", "automation"], "updated_at": "2024-01-07", "url": "https://arxiv.org/abs/2304.07061"}
{"aliases": [], "authors": ["Yihong Dong", "Xue Jiang", "Zhi Jin", "Ge Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "30e6371864ee30ac78378cc25640b8f9336e5162f984fa44cf652a5f50ecfb51", "submitted_at": "2023-04-15T16:33:32Z", "updated_at": "2024-05-11T14:00:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2304.07590v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2304.07590", "identifiers": {"arxiv": "2304.07590"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2304.07590"}], "name": "Self-collaboration Code Generation via ChatGPT", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-04-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-11", "url": "https://arxiv.org/abs/2304.07590"}
{"aliases": [], "authors": ["Yujia Qin", "Shengding Hu", "Yankai Lin", "Weize Chen", "Ning Ding", "Ganqu Cui", "Zheni Zeng", "Yufei Huang", "Chaojun Xiao", "Chi Han", "Yi Ren Fung", "Yusheng Su", "Huadong Wang", "Cheng Qian", "Runchu Tian", "Kunlun Zhu", "Shihao Liang", "Xingyu Shen", "Bokai Xu", "Zhen Zhang", "Yining Ye", "Bowen Li", "Ziwei Tang", "Jing Yi", "Yuzhang Zhu", "Zhenning Dai", "Lan Yan", "Xin Cong", "Yaxi Lu", "Weilin Zhao", "Yuxiang Huang", "Junxi Yan", "Xu Han", "Xian Sun", "Dahai Li", "Jason Phang", "Cheng Yang", "Tongshuang Wu", "Heng Ji", "Zhiyuan Liu", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9987288e7dfa82bca811599beb8880732f90ea971ea15676e194360cbdd2a32e", "submitted_at": "2023-04-17T15:16:10Z", "updated_at": "2024-08-06T15:14:36Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2304.08354v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2304.08354", "identifiers": {"arxiv": "2304.08354"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2304.08354"}], "name": "Tool Learning with Foundation Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-04-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-06", "url": "https://arxiv.org/abs/2304.08354"}
{"aliases": [], "authors": ["Jindong Gu", "Ahmad Beirami", "Xuezhi Wang", "Alex Beutel", "Philip Torr", "Yao Qin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "875f73a8117cc69d4b15a57fdc27d981aaba7e71dd956fbb7a0686ecfb52d251", "submitted_at": "2023-04-17T17:58:07Z", "updated_at": "2023-04-17T17:58:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2304.08479v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2304.08479", "identifiers": {"arxiv": "2304.08479"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2304.08479"}], "name": "Towards Robust Prompts on Vision-Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-04-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-04-17", "url": "https://arxiv.org/abs/2304.08479"}
{"aliases": [], "authors": ["Ankit Shukla", "Laura Heller", "Md Golam Morshed", "Laura Rehm", "Avik W. Ghosh", "Andrew D. Kent", "Shaloo Rakheja"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "008aed6959b4bf8ebda80c63cea9857b8c6ffa5d3f566dfdbcaed070550bbc31", "submitted_at": "2023-04-18T08:17:31Z", "updated_at": "2023-04-18T08:17:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2304.08808v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2304.08808", "identifiers": {"arxiv": "2304.08808"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2304.08808"}], "name": "A True Random Number Generator for Probabilistic Computing using Stochastic Magnetic Actuated Random Transducer Devices", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-04-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-04-18", "url": "https://arxiv.org/abs/2304.08808"}
{"aliases": [], "authors": ["Murat Isik", "Kayode Inadagbo", "Hakan Aktas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "adae946c8b48280b5af191cce824df9187cb04aa48a2e1d821d091be1d54ebd1", "submitted_at": "2023-04-24T22:20:42Z", "updated_at": "2023-04-24T22:20:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2304.12474v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2304.12474", "identifiers": {"arxiv": "2304.12474"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2304.12474"}], "name": "Design optimization for high-performance computing using FPGA", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-04-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-04-24", "url": "https://arxiv.org/abs/2304.12474"}
{"aliases": [], "authors": ["Alexander Wan", "Eric Wallace", "Sheng Shen", "Dan Klein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "281fb97b8f5b0c15d9afe4a22768cbb6e43e67929fd6ab151f81df8f8ede648f", "submitted_at": "2023-05-01T16:57:33Z", "updated_at": "2023-05-01T16:57:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.00944v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2305.00944", "identifiers": {"arxiv": "2305.00944"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.00944"}], "name": "Poisoning Language Models During Instruction Tuning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-01", "url": "https://arxiv.org/abs/2305.00944"}
{"aliases": [], "authors": ["Suryansh Upadhyay", "Swaroop Ghosh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40ed53b48862c09a70992d60b1496e214f957a6ca9be950527cbbba6065719e9", "submitted_at": "2023-05-02T23:34:01Z", "updated_at": "2024-11-14T20:47:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.01826v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2305.01826", "identifiers": {"arxiv": "2305.01826"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.01826"}], "name": "Trustworthy Computing using Untrusted Cloud-Based Quantum Hardware", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-14", "url": "https://arxiv.org/abs/2305.01826"}
{"aliases": [], "authors": ["Lichang Chen", "Minhao Cheng", "Heng Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fdde236a5c8635529a6a7531ae6b18409cc7131ae522a42200bf05784b9f43ee", "submitted_at": "2023-05-03T20:31:13Z", "updated_at": "2023-05-03T20:31:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.02424v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2305.02424", "identifiers": {"arxiv": "2305.02424"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.02424"}], "name": "Backdoor Learning on Sequence to Sequence Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-03", "url": "https://arxiv.org/abs/2305.02424"}
{"aliases": [], "authors": ["Lei Wang", "Wanyu Xu", "Yihuai Lan", "Zhiqiang Hu", "Yunshi Lan", "Roy Ka-Wei Lee", "Ee-Peng Lim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3dc514e909b703f3f24440ca95d0ef3fe82285e17afd77a1bb2ec7b5c5a769b8", "submitted_at": "2023-05-06T16:34:37Z", "updated_at": "2023-05-26T07:06:48Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2305.04091v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2305.04091", "identifiers": {"arxiv": "2305.04091"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.04091"}], "name": "Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-26", "url": "https://arxiv.org/abs/2305.04091"}
{"aliases": [], "authors": ["Zhaoyang Liu", "Yinan He", "Wenhai Wang", "Weiyun Wang", "Yi Wang", "Shoufa Chen", "Qinglong Zhang", "Zeqiang Lai", "Yang Yang", "Qingyun Li", "Jiashuo Yu", "Kunchang Li", "Zhe Chen", "Xue Yang", "Xizhou Zhu", "Yali Wang", "Limin Wang", "Ping Luo", "Jifeng Dai", "Yu Qiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae5d22dd06ac06905b43804d3f654f2e34b4d5dc31aa8abc478f8c1fdce4a18b", "submitted_at": "2023-05-09T17:58:34Z", "updated_at": "2023-06-02T16:19:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2305.05662v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2305.05662", "identifiers": {"arxiv": "2305.05662"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.05662"}], "name": "InternGPT: Solving Vision-Centric Tasks by Interacting with ChatGPT Beyond Language", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-06-02", "url": "https://arxiv.org/abs/2305.05662"}
{"aliases": [], "authors": ["René Bødker Christensen", "Petar Popovski"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "488909fbd51001b1e656d1376261595666f30253aca115a994b998d82ca28615", "submitted_at": "2023-05-10T09:01:38Z", "updated_at": "2023-10-27T09:50:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.05993v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2305.05993", "identifiers": {"arxiv": "2305.05993"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.05993"}], "name": "Private Product Computation using Quantum Entanglement", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-27", "url": "https://arxiv.org/abs/2305.05993"}
{"aliases": [], "authors": ["Danyang Zhang", "Zhennan Shen", "Rui Xie", "Situo Zhang", "Tianbao Xie", "Zihan Zhao", "Siyuan Chen", "Lu Chen", "Hongshen Xu", "Ruisheng Cao", "Kai Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e454f547bdf8e01f47eb63c20f0cacfa8bd17a8c4ad37dfd3a1cceb732de178", "submitted_at": "2023-05-14T12:31:03Z", "updated_at": "2024-06-13T11:51:37Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2305.08144v4"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2305.08144", "identifiers": {"arxiv": "2305.08144"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2305.08144"}], "name": "Mobile-Env: Building Qualified Evaluation Benchmarks for LLM-GUI Interaction", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-05-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Provides an Android benchmark toolkit with controlled environments, intermediate feedback, and task sets for reproducible GUI-agent evaluation.", "topics": ["evaluation", "sandbox"], "updated_at": "2024-06-13", "url": "https://arxiv.org/abs/2305.08144"}
{"aliases": [], "authors": ["Zhangxuan Gu", "Zhuoer Xu", "Haoxing Chen", "Jun Lan", "Changhua Meng", "Weiqiang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cbd23cad847d29812a6adca85eedfafb4a55bb33395d6e9907e34b8cd1218745", "submitted_at": "2023-05-16T07:16:36Z", "updated_at": "2023-05-16T07:16:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.09699v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2305.09699", "identifiers": {"arxiv": "2305.09699"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.09699"}], "name": "Mobile User Interface Element Detection Via Adaptively Prompt Tuning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-16", "url": "https://arxiv.org/abs/2305.09699"}
{"aliases": [], "authors": ["Hua Shen", "Chieh-Yang Huang", "Tongshuang Wu", "Ting-Hao 'Kenneth' Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "025ca9d228de6d7513fe5fb4893b750b0bc05b5bd2ea3cad87228fb2884798c8", "submitted_at": "2023-05-16T19:48:49Z", "updated_at": "2023-10-27T16:08:32Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2305.09770v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2305.09770", "identifiers": {"arxiv": "2305.09770"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.09770"}], "name": "ConvXAI: Delivering Heterogeneous AI Explanations via Conversations to Support Human-AI Scientific Writing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-27", "url": "https://arxiv.org/abs/2305.09770"}
{"aliases": [], "authors": ["Shunyu Yao", "Dian Yu", "Jeffrey Zhao", "Izhak Shafran", "Thomas L. Griffiths", "Yuan Cao", "Karthik Narasimhan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0ab6699333caf8dc49aee0ea63a40a87e4d857489191f8d88d7dd38a8939ed9", "submitted_at": "2023-05-17T23:16:17Z", "updated_at": "2023-12-03T22:50:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.10601v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2305.10601", "identifiers": {"arxiv": "2305.10601"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.10601"}], "name": "Tree of Thoughts: Deliberate Problem Solving with Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-03", "url": "https://arxiv.org/abs/2305.10601"}
{"aliases": [], "authors": ["Ning Lu", "Shengcai Liu", "Rui He", "Qi Wang", "Yew-Soon Ong", "Ke Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f03e6fd3f0ac728fe86ff0eaf36e849c389dba4a071d11a6d204709a647f6cc0", "submitted_at": "2023-05-18T10:03:25Z", "updated_at": "2024-05-15T08:00:09Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2305.10847v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2305.10847", "identifiers": {"arxiv": "2305.10847"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.10847"}], "name": "Large Language Models can be Guided to Evade AI-Generated Text Detection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-15", "url": "https://arxiv.org/abs/2305.10847"}
{"aliases": [], "authors": ["Cristian-Paul Bara", "Ziqiao Ma", "Yingzhuo Yu", "Julie Shah", "Joyce Chai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c462e2b191256ee69e8e929495462075dcae1b092f4d9b8d579f5b56bed6d7f8", "submitted_at": "2023-05-18T19:42:04Z", "updated_at": "2023-05-18T19:42:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.11271v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2305.11271", "identifiers": {"arxiv": "2305.11271"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.11271"}], "name": "Towards Collaborative Plan Acquisition through Theory of Mind Modeling in Situated Dialogue", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-18", "url": "https://arxiv.org/abs/2305.11271"}
{"aliases": [], "authors": ["Hiroki Furuta", "Kuang-Huei Lee", "Ofir Nachum", "Yutaka Matsuo", "Aleksandra Faust", "Shixiang Shane Gu", "Izzeddin Gur"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac0ea5a462890317ea29ade32c879d016181afbe5962e2d28cc175e251df41e6", "submitted_at": "2023-05-19T17:44:34Z", "updated_at": "2024-02-25T16:21:00Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2305.11854v4"}, "publication": {"authors": ["Hiroki Furuta", "Kuang-Huei Lee", "Ofir Nachum", "Yutaka Matsuo", "Aleksandra Faust", "Shixiang Shane Gu", "Izzeddin Gur"], "date_basis": "Explicit publisher metadata: 2024-05-31; precision=day.", "date_parts": [2024, 5, 31], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html", "https://proceedings.iclr.cc/paper_files/paper/2024/file/7ef7d8359036afd8c2378d82c21058a4-Paper-Conference.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The ICLR landing shortens Shixiang Shane Gu to Shixiang Gu. Publication authors use the full PDF title-page form. Volume 2024 and pages 29528–29558 are explicit publisher citation metadata, not values inferred from the event year or PDF page count. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "29528-29558", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2024/file/7ef7d8359036afd8c2378d82c21058a4-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Multimodal Web Navigation with Instruction-Finetuned Foundation Models", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2024"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "6669190c7fc2a37f9cd68d0c49431b8f17b626c16bb3ef4093fc83a6ac62acd3", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:46+00:00", "role": "primary", "sha256": "da26ea085252edfda4c084058aa81ea023f885ff97c67f2ffae0698bab6f568b", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://proceedings.iclr.cc/paper_files/paper/2024/file/7ef7d8359036afd8c2378d82c21058a4-Paper-Conference.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2305.11854", "identifiers": {"arxiv": "2305.11854"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2305.11854"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2024/file/7ef7d8359036afd8c2378d82c21058a4-Paper-Conference.pdf"}], "name": "Multimodal Web Navigation with Instruction-Finetuned Foundation Models", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2023-05-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains WebGUM offline from demonstrations by jointly tuning an instruction-following language model and a vision encoder over webpage screenshots and HTML. The agent predicts browser actions and is evaluated on MiniWoB, WebShop, and Mind2Web.", "topics": ["web-navigation", "imitation-learning", "action-prediction"], "updated_at": "2024-02-25", "url": "https://arxiv.org/abs/2305.11854"}
{"aliases": [], "authors": ["Cédric Colas", "Laetitia Teodorescu", "Pierre-Yves Oudeyer", "Xingdi Yuan", "Marc-Alexandre Côté"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "444a44752af982de9b0147c891e31dc71f9733ddffba862340f09455fe87d272", "submitted_at": "2023-05-21T15:42:41Z", "updated_at": "2023-05-21T15:42:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.12487v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2305.12487", "identifiers": {"arxiv": "2305.12487"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.12487"}], "name": "Augmenting Autotelic Agents with Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-21", "url": "https://arxiv.org/abs/2305.12487"}
{"aliases": [], "authors": ["Jian Xie", "Kai Zhang", "Jiangjie Chen", "Renze Lou", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "92b4bb894fcc6c9d3a0a0b274b37eaf3b4595ea41301c6a6e40a36c2924c6c59", "submitted_at": "2023-05-22T17:57:41Z", "updated_at": "2024-02-27T17:08:49Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2305.13300v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2305.13300", "identifiers": {"arxiv": "2305.13300"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.13300"}], "name": "Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-27", "url": "https://arxiv.org/abs/2305.13300"}
{"aliases": [], "authors": ["Abishek Sridhar", "Robert Lo", "Frank F. Xu", "Hao Zhu", "Shuyan Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "896d5f5e9248e2e49da254fb383f84e5b6ce4286b6d30a37ed7944fcfa1f4958", "submitted_at": "2023-05-23T17:10:39Z", "updated_at": "2023-10-30T00:55:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2305.14257v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2305.14257", "identifiers": {"arxiv": "2305.14257"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2305.14257"}], "name": "Hierarchical Prompting Assists Large Language Model on Web Navigation", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2023-05-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates web navigation prompting into a SUMMARIZER that extracts action-relevant observations and an ACTOR that chooses the next action from the condensed context.", "topics": ["web-navigation", "planning", "efficiency"], "updated_at": "2023-10-30", "url": "https://arxiv.org/abs/2305.14257"}
{"aliases": [], "authors": ["Yilun Du", "Shuang Li", "Antonio Torralba", "Joshua B. Tenenbaum", "Igor Mordatch"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "297c58d2a4ad1af7b95985541bf1b6a33b0f476910e807b4fb709ac6f29df5e2", "submitted_at": "2023-05-23T17:55:11Z", "updated_at": "2023-05-23T17:55:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.14325v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2305.14325", "identifiers": {"arxiv": "2305.14325"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.14325"}], "name": "Improving Factuality and Reasoning in Language Models through Multiagent Debate", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-23", "url": "https://arxiv.org/abs/2305.14325"}
{"aliases": [], "authors": ["Alejandro Escontrela", "Ademi Adeniji", "Wilson Yan", "Ajay Jain", "Xue Bin Peng", "Ken Goldberg", "Youngwoon Lee", "Danijar Hafner", "Pieter Abbeel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1220719adbf421bc55570c5f792481e69daaa7dd5b886d84eeba363365ffbaf8", "submitted_at": "2023-05-23T17:59:33Z", "updated_at": "2023-05-30T17:38:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.14343v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2305.14343", "identifiers": {"arxiv": "2305.14343"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.14343"}], "name": "Video Prediction Models as Rewards for Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-30", "url": "https://arxiv.org/abs/2305.14343"}
{"aliases": [], "authors": ["Jiashu Xu", "Mingyu Derek Ma", "Fei Wang", "Chaowei Xiao", "Muhao Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ee440d1d79a2a30244c5c9923633adf6a7c5d82ad5087e5548c68ed383131b63", "submitted_at": "2023-05-24T04:27:21Z", "updated_at": "2024-04-03T09:15:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.14710v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2305.14710", "identifiers": {"arxiv": "2305.14710"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.14710"}], "name": "Instructions as Backdoors: Backdoor Vulnerabilities of Instruction Tuning for Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-03", "url": "https://arxiv.org/abs/2305.14710"}
{"aliases": [], "authors": ["Jiongxiao Wang", "Zichen Liu", "Keun Hee Park", "Zhuojun Jiang", "Zhaoheng Zheng", "Zhuofeng Wu", "Muhao Chen", "Chaowei Xiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e131fe72d21e101568ee759c0236a4db5201ae14a0bb34ad80fd53071d0066f5", "submitted_at": "2023-05-24T09:40:56Z", "updated_at": "2023-10-14T05:03:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.14950v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2305.14950", "identifiers": {"arxiv": "2305.14950"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.14950"}], "name": "Adversarial Demonstration Attacks on Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-14", "url": "https://arxiv.org/abs/2305.14950"}
{"aliases": [], "authors": ["Shishir G. Patil", "Tianjun Zhang", "Xin Wang", "Joseph E. Gonzalez"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e03b219fe6ddfb7556e0b4482c8efc8cb990a5813c3999745b8a0201f0d9cec", "submitted_at": "2023-05-24T16:48:11Z", "updated_at": "2023-05-24T16:48:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.15334v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2305.15334", "identifiers": {"arxiv": "2305.15334"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.15334"}], "name": "Gorilla: Large Language Model Connected with Massive APIs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-24", "url": "https://arxiv.org/abs/2305.15334"}
{"aliases": [], "authors": ["Yue Wu", "Shrimai Prabhumoye", "So Yeon Min", "Yonatan Bisk", "Ruslan Salakhutdinov", "Amos Azaria", "Tom Mitchell", "Yuanzhi Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b88ce1b9dd7c87bc27efa874466729922b9121e62e910cf5fa83469ed5c22dbe", "submitted_at": "2023-05-24T18:14:35Z", "updated_at": "2023-12-11T22:18:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2305.15486v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2305.15486", "identifiers": {"arxiv": "2305.15486"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.15486"}], "name": "SPRING: Studying the Paper and Reasoning to Play Games", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-11", "url": "https://arxiv.org/abs/2305.15486"}
{"aliases": [], "authors": ["Mengyu Zhang", "Xiangyu Ren", "Guanglei Xi", "Zhenxing Zhang", "Qiaonian Yu", "Fuming Liu", "Hualiang Zhang", "Shengyu Zhang", "Yi-Cong Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "337989043423afbaf3e05c86ea3772b9498630c12b1fdce87d87b06510e430fd", "submitted_at": "2023-05-25T06:23:32Z", "updated_at": "2023-05-25T06:23:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.15767v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2305.15767", "identifiers": {"arxiv": "2305.15767"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.15767"}], "name": "A Scalable, Fast and Programmable Neural Decoder for Fault-Tolerant Quantum Computation Using Surface Codes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-25", "url": "https://arxiv.org/abs/2305.15767"}
{"aliases": [], "authors": ["Guanzhi Wang", "Yuqi Xie", "Yunfan Jiang", "Ajay Mandlekar", "Chaowei Xiao", "Yuke Zhu", "Linxi Fan", "Anima Anandkumar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86d47d4e1f2fdaa5a14c1efbc5c3d5986169f6c6a1e562b3d0ffff68154dbd64", "submitted_at": "2023-05-25T17:46:38Z", "updated_at": "2023-10-19T16:27:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.16291v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2305.16291", "identifiers": {"arxiv": "2305.16291"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.16291"}], "name": "Voyager: An Open-Ended Embodied Agent with Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-19", "url": "https://arxiv.org/abs/2305.16291"}
{"aliases": [], "authors": ["Jikun Kang", "Romain Laroche", "Xingdi Yuan", "Adam Trischler", "Xue Liu", "Jie Fu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "93ab0fb88e9b1cf0e929085239db21884aa23fad732ed618219fcb0622ddebad", "submitted_at": "2023-05-24T01:20:22Z", "updated_at": "2024-05-28T19:03:33Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2305.16338v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2305.16338", "identifiers": {"arxiv": "2305.16338"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.16338"}], "name": "Think Before You Act: Decision Transformers with Working Memory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-28", "url": "https://arxiv.org/abs/2305.16338"}
{"aliases": [], "authors": ["Qiantong Xu", "Fenglu Hong", "Bo Li", "Changran Hu", "Zhengyu Chen", "Jian Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "819b5bf1cef7b8f674cab337c972e1fd365ffc3f188cf4c169d55a3b5e692593", "submitted_at": "2023-05-25T22:10:20Z", "updated_at": "2023-05-25T22:10:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.16504v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2305.16504", "identifiers": {"arxiv": "2305.16504"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.16504"}], "name": "On the Tool Manipulation Capability of Open-source Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-25", "url": "https://arxiv.org/abs/2305.16504"}
{"aliases": [], "authors": ["Haotian Sun", "Yuchen Zhuang", "Lingkai Kong", "Bo Dai", "Chao Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c254a26ebb8e1bcdca70b7581c780aa6b0dd4e5f5ded85e456f7e8fb8b38fcc6", "submitted_at": "2023-05-26T05:52:27Z", "updated_at": "2023-05-26T05:52:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.16653v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2305.16653", "identifiers": {"arxiv": "2305.16653"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2305.16653"}], "name": "AdaPlanner: Adaptive Planning from Feedback with Language Models", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. This is a general planning framework; MiniWoB++ supplies the interface-agent evidence, while ALFWorld is a separate environment.", "published_at": "2023-05-26", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces AdaPlanner, which revises code-structured plans in response to environmental feedback and reuses successful plans as exemplars. Experiments include MiniWoB++ web tasks alongside ALFWorld.", "topics": ["planning", "self-reflection", "evaluation"], "updated_at": "2023-05-26", "url": "https://arxiv.org/abs/2305.16653"}
{"aliases": [], "authors": ["David Brandfonbrener", "Ofir Nachum", "Joan Bruna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a7f5b21507dbd2f4e0e03b409ce54eab7189a71420aaf3bb8d8a43c910a5796", "submitted_at": "2023-05-26T14:40:46Z", "updated_at": "2023-10-25T17:59:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.16985v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2305.16985", "identifiers": {"arxiv": "2305.16985"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.16985"}], "name": "Inverse Dynamics Pretraining Learns Good Representations for Multitask Imitation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-25", "url": "https://arxiv.org/abs/2305.16985"}
{"aliases": [], "authors": ["Xizhou Zhu", "Yuntao Chen", "Hao Tian", "Chenxin Tao", "Weijie Su", "Chenyu Yang", "Gao Huang", "Bin Li", "Lewei Lu", "Xiaogang Wang", "Yu Qiao", "Zhaoxiang Zhang", "Jifeng Dai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e296992e6535846900a2651212d6d00e4db8427293d1dadad75ae6f4d7d47a0", "submitted_at": "2023-05-25T17:59:49Z", "updated_at": "2023-06-01T09:18:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.17144v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2305.17144", "identifiers": {"arxiv": "2305.17144"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.17144"}], "name": "Ghost in the Minecraft: Generally Capable Agents for Open-World Environments via Large Language Models with Text-based Knowledge and Memory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-06-01", "url": "https://arxiv.org/abs/2305.17144"}
{"aliases": [], "authors": ["John P. T. Stenger", "C. Stephen Hellberg", "Daniel Gunlycke"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4fb07ca894b1c33070365ccf2ee8b75c8968a1e3fad6584a5e32533894e62aab", "submitted_at": "2023-05-30T13:10:44Z", "updated_at": "2023-05-30T13:10:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.19014v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2305.19014", "identifiers": {"arxiv": "2305.19014"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2305.19014"}], "name": "Implementing Jastrow--Gutzwiller operators on a quantum computer using the cascaded variational quantum eigensolver algorithm", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-05-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-05-30", "url": "https://arxiv.org/abs/2305.19014"}
{"aliases": [], "authors": ["Hongxin Li", "Jingran Su", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "320ff40872677e9d32efae360dc289ca345076e037e7044188868f765e816008", "submitted_at": "2023-05-30T17:59:30Z", "updated_at": "2023-10-30T06:36:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.19308v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:56Z", "role": "primary", "sha256": "beb8c811bf2d070845aa6fcc7e94ffd43ca4f9f60ce42c72d958598790fbdffc", "supports": ["environments", "notes"], "url": "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/README.md"}], "id": "arxiv:2305.19308", "identifiers": {"arxiv": "2305.19308"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2305.19308"}], "name": "SheetCopilot: Bringing Software Productivity to the Next Level through Large Language Models", "notes": "The contribution was reviewed from the retained paper abstract. The original Windows/Excel implementation is identified in the separately retained project README; the later Ubuntu evaluator is not treated as a paper result. No results were independently reproduced.", "published_at": "2023-05-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SheetCopilot, an agent that steers spreadsheet software via atomic virtual APIs, along with a 221-task spreadsheet manipulation benchmark.", "topics": ["workflow", "tool-integration", "evaluation"], "updated_at": "2023-10-30", "url": "https://arxiv.org/abs/2305.19308"}
{"aliases": [], "authors": ["Peter Shaw", "Mandar Joshi", "James Cohan", "Jonathan Berant", "Panupong Pasupat", "Hexiang Hu", "Urvashi Khandelwal", "Kenton Lee", "Kristina Toutanova"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d06f7480695f45546b62b5b8c3d8c174ca52936e158d46e626689d0918711417", "submitted_at": "2023-05-31T23:39:18Z", "updated_at": "2023-12-06T23:46:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.00245v2"}, "publication": {"authors": ["Peter Shaw", "Mandar Joshi", "James Cohan", "Jonathan Berant", "Panupong Pasupat", "Hexiang Hu", "Urvashi Khandelwal", "Kenton Lee", "Kristina Toutanova"], "date_basis": "Explicit publisher metadata: 2023-12-15; precision=day.", "date_parts": [2023, 12, 15], "doi": "10.52202/075280-1490", "event_dates": null, "evidence_urls": ["https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html", "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The NeurIPS landing metadata adds the initial N to Kristina Toutanova; the formal paper title page does not. Publication authors use the PDF. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "34354-34370", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf", "publisher": null, "series": null, "title": "From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces", "type": "conference-paper", "url": "https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html", "venue": "Advances in Neural Information Processing Systems", "volume": "36"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "938e0a31512bb9a7793b568858f81eb3ceb650a71aae2d91b5292990e890201b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:47+00:00", "role": "primary", "sha256": "3827298f44d2f24f1333b4d46ecb5eb88c8b028d192b11424580655b38e024f9", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2306.00245", "identifiers": {"arxiv": "2306.00245"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2306.00245"}, {"rel": "publication", "url": "https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf"}], "name": "From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2023-05-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies instruction-following agents that observe screenshots and act through a generic mouse-and-keyboard action space, evaluating pixel-based pretraining for GUI tasks on MiniWoB++.", "topics": ["grounding", "action-prediction", "web-navigation"], "updated_at": "2023-12-06", "url": "https://arxiv.org/abs/2306.00245"}
{"aliases": [], "authors": ["Alp Sahin", "Subhrajit Bhattacharya"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "99729608976eec31fb82286f7de82738f12fa7d809e3a71d5057adedd9cf7c0a", "submitted_at": "2023-06-01T23:33:53Z", "updated_at": "2023-07-10T21:55:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.01203v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2306.01203", "identifiers": {"arxiv": "2306.01203"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.01203"}], "name": "Topo-Geometrically Distinct Path Computation using Neighborhood-augmented Graph, and its Application to Path Planning for a Tethered Robot in 3D", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-07-10", "url": "https://arxiv.org/abs/2306.01203"}
{"aliases": [], "authors": ["Zhizheng Zhang", "Xiaoyi Zhang", "Wenxuan Xie", "Yan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "631a39d2929c68bb93a7f95af22c17d2ab208668fd97d9082fe0601a73d123dd", "submitted_at": "2023-06-02T02:42:58Z", "updated_at": "2023-12-04T13:36:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.01242v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2306.01242", "identifiers": {"arxiv": "2306.01242"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2306.01242"}], "name": "Responsible Task Automation: Empowering Large Language Models as Responsible Task Automators", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. UI task automation is stated in the abstract, but its platform and detailed evaluation protocol are not identified.", "published_at": "2023-06-02", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Presents ResponsibleTA, a coordinator-executor framework that checks command feasibility and task completion while using local memory for privacy-related controls. The paper compares prompted and learned implementations and evaluates them on UI task automation.", "topics": ["safety-security", "evaluation", "automation"], "updated_at": "2023-12-04", "url": "https://arxiv.org/abs/2306.01242"}
{"aliases": [], "authors": ["Florent De Geeter", "Damien Ernst", "Guillaume Drion"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d660e7c8ddf6f5f8663f07ece23b73ea2fd4605c8a5a39ee74c87f5bf8dc9fde", "submitted_at": "2023-06-06T12:19:12Z", "updated_at": "2024-05-06T12:08:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2306.03623v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2306.03623", "identifiers": {"arxiv": "2306.03623"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.03623"}], "name": "Spike-based computation using classical recurrent neural networks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-06", "url": "https://arxiv.org/abs/2306.03623"}
{"aliases": [], "authors": ["Joseph Harris", "Elham Kashefi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac195dda5e673407767878aa5c1cabfc65783aa82aac41fb748c8a49f366516a", "submitted_at": "2023-06-07T11:35:21Z", "updated_at": "2025-02-05T09:43:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2306.04351v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2306.04351", "identifiers": {"arxiv": "2306.04351"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.04351"}], "name": "Error Mitigation of BQP Computations using Measurement-Based Verification", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-05", "url": "https://arxiv.org/abs/2306.04351"}
{"aliases": [], "authors": ["Nikhil Kandpal", "Brian Lester", "Mohammed Muqeeth", "Anisha Mascarenhas", "Monty Evans", "Vishal Baskaran", "Tenghao Huang", "Haokun Liu", "Colin Raffel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "588b92153d4fe7edf14b8a2e39d98b8f0c8ab945262a97368aba0098c2c568ad", "submitted_at": "2023-06-07T15:37:50Z", "updated_at": "2023-06-07T15:37:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2306.04529v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2306.04529", "identifiers": {"arxiv": "2306.04529"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.04529"}], "name": "Git-Theta: A Git Extension for Collaborative Development of Machine Learning Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-06-07", "url": "https://arxiv.org/abs/2306.04529"}
{"aliases": [], "authors": ["Yi Liu", "Gelei Deng", "Yuekang Li", "Kailong Wang", "Zihao Wang", "Xiaofeng Wang", "Tianwei Zhang", "Yepang Liu", "Haoyu Wang", "Yan Zheng", "Leo Yu Zhang", "Yang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "556427a8251d31f12c635e72e9e3092a88e3e1f24bbb8e83e664a9853f068ad6", "submitted_at": "2023-06-08T18:43:11Z", "updated_at": "2025-12-29T02:25:27Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2306.05499v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2306.05499", "identifiers": {"arxiv": "2306.05499"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.05499"}], "name": "Prompt Injection attack against LLM-integrated Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-29", "url": "https://arxiv.org/abs/2306.05499"}
{"aliases": [], "authors": ["Xiang Deng", "Yu Gu", "Boyuan Zheng", "Shijie Chen", "Samuel Stevens", "Boshi Wang", "Huan Sun", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de055dfa452130f3576839c99c007053c62a0c96c9bfbcdd79ecad1a9f74f830", "submitted_at": "2023-06-09T17:44:31Z", "updated_at": "2023-12-09T05:57:46Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2306.06070v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2306.06070", "identifiers": {"arxiv": "2306.06070"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2306.06070"}], "name": "Mind2Web: Towards a Generalist Agent for the Web", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2023-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a web-task dataset spanning diverse websites and evaluates generalization to unseen sites and domains.", "topics": ["data-generation", "web-navigation", "evaluation"], "updated_at": "2023-12-09", "url": "https://arxiv.org/abs/2306.06070"}
{"aliases": [], "authors": ["Yifan Song", "Weimin Xiong", "Dawei Zhu", "Wenhao Wu", "Han Qian", "Mingbo Song", "Hailiang Huang", "Cheng Li", "Ke Wang", "Rong Yao", "Ye Tian", "Sujian Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "83409092e358d43533095bc4a372a5544c44a5635694dde7729efc520c5f0543", "submitted_at": "2023-06-11T08:53:12Z", "updated_at": "2023-08-27T02:55:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.06624v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2306.06624", "identifiers": {"arxiv": "2306.06624"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.06624"}], "name": "RestGPT: Connecting Large Language Models with Real-World RESTful APIs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-27", "url": "https://arxiv.org/abs/2306.06624"}
{"aliases": [], "authors": ["George E. Dahl", "Frank Schneider", "Zachary Nado", "Naman Agarwal", "Chandramouli Shama Sastry", "Philipp Hennig", "Sourabh Medapati", "Runa Eschenhagen", "Priya Kasimbeg", "Daniel Suo", "Juhan Bae", "Justin Gilmer", "Abel L. Peirson", "Bilal Khan", "Rohan Anil", "Mike Rabbat", "Shankar Krishnan", "Daniel Snider", "Ehsan Amid", "Kongtao Chen", "Chris J. Maddison", "Rakshith Vasudev", "Michal Badura", "Ankush Garg", "Peter Mattson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fef5617bbc1e5bd3f412871c93e8b80843ab4463240ffd34d07b036856ede9e1", "submitted_at": "2023-06-12T15:21:02Z", "updated_at": "2025-06-18T11:00:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.07179v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2306.07179", "identifiers": {"arxiv": "2306.07179"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.07179"}], "name": "Benchmarking Neural Network Training Algorithms", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-18", "url": "https://arxiv.org/abs/2306.07179"}
{"aliases": [], "authors": ["Longtao Zheng", "Rundong Wang", "Xinrun Wang", "Bo An"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8fdb5fed2b2fa0ecbf5910dd6813d8de7910af286a1e4e00cdebb734a82e78b0", "submitted_at": "2023-06-13T15:49:41Z", "updated_at": "2024-01-19T06:59:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2306.07863v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2306.07863", "identifiers": {"arxiv": "2306.07863"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2306.07863"}], "name": "Synapse: Trajectory-as-Exemplar Prompting with Memory for Computer Control", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-06-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Synapse filters task-irrelevant page state and places complete abstracted state-action trajectories into the prompt. An embedding-based exemplar memory retrieves relevant demonstrations for new tasks, evaluated on MiniWoB++ and Mind2Web.", "topics": ["memory", "web-navigation", "planning", "efficiency"], "updated_at": "2024-01-19", "url": "https://arxiv.org/abs/2306.07863"}
{"aliases": [], "authors": ["Ivan S. Maksymov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c76bea480a1764932e51083205b5513b52ee220fdbe9cd4effea227585c2cc4d", "submitted_at": "2023-06-15T12:47:22Z", "updated_at": "2023-06-15T12:47:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2306.09095v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2306.09095", "identifiers": {"arxiv": "2306.09095"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.09095"}], "name": "Analogue and Physical Reservoir Computing Using Water Waves", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-06-15", "url": "https://arxiv.org/abs/2306.09095"}
{"aliases": [], "authors": ["Shuyuan Xu", "Wenyue Hua", "Yongfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "71a97724462f60a55159087d5fbe43befffca0b4a5055dc9fab5423a522ab5e4", "submitted_at": "2023-06-19T19:38:29Z", "updated_at": "2024-04-11T02:42:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.11134v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2306.11134", "identifiers": {"arxiv": "2306.11134"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.11134"}], "name": "OpenP5: An Open-Source Platform for Developing, Training, and Evaluating LLM-based Recommender Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-11", "url": "https://arxiv.org/abs/2306.11134"}
{"aliases": [], "authors": ["Xiangyu Qi", "Kaixuan Huang", "Ashwinee Panda", "Peter Henderson", "Mengdi Wang", "Prateek Mittal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1e610e5a98c7a94fb342ff1775f50879138b9732d73b22b1a0185c30bdd2fc78", "submitted_at": "2023-06-22T22:13:03Z", "updated_at": "2023-08-16T22:38:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.13213v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2306.13213", "identifiers": {"arxiv": "2306.13213"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.13213"}], "name": "Visual Adversarial Examples Jailbreak Aligned Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-16", "url": "https://arxiv.org/abs/2306.13213"}
{"aliases": [], "authors": ["Nicholas Carlini", "Milad Nasr", "Christopher A. Choquette-Choo", "Matthew Jagielski", "Irena Gao", "Anas Awadalla", "Pang Wei Koh", "Daphne Ippolito", "Katherine Lee", "Florian Tramer", "Ludwig Schmidt"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b20d95ee3db73e68f6ec5e876a40bcb908abf77ba3afccc3adf0ebd12373d74", "submitted_at": "2023-06-26T17:18:44Z", "updated_at": "2024-05-06T06:36:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.15447v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2306.15447", "identifiers": {"arxiv": "2306.15447"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.15447"}], "name": "Are aligned neural networks adversarially aligned?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-06", "url": "https://arxiv.org/abs/2306.15447"}
{"aliases": [], "authors": ["Manli Shu", "Jiongxiao Wang", "Chen Zhu", "Jonas Geiping", "Chaowei Xiao", "Tom Goldstein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "510e060f32d21c5f51974129e19c3581c676a53f9c83a21b72bec1cc84538113", "submitted_at": "2023-06-28T17:54:04Z", "updated_at": "2023-10-28T18:04:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.17194v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2306.17194", "identifiers": {"arxiv": "2306.17194"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2306.17194"}], "name": "On the Exploitability of Instruction Tuning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-06-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-28", "url": "https://arxiv.org/abs/2306.17194"}
{"aliases": [], "authors": ["Zhongjin Luo", "Dong Du", "Heming Zhu", "Yizhou Yu", "Hongbo Fu", "Xiaoguang Han"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d74e58acea23cf46d1e49808561a4614c421c5240e30fe3a930d455f544d94cb", "submitted_at": "2023-07-03T07:41:07Z", "updated_at": "2023-07-04T12:21:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2307.00804v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2307.00804", "identifiers": {"arxiv": "2307.00804"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2307.00804"}], "name": "SketchMetaFace: A Learning-based Sketching Interface for High-fidelity 3D Character Face Modeling", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-07-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-07-04", "url": "https://arxiv.org/abs/2307.00804"}
{"aliases": [], "authors": ["Dominic Gold", "Koray Karabina", "Francis C. Motta"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "620d8bd887eaa8d4543b260cbb4fc9019c5a35632d74f0488c22bf2ff105a930", "submitted_at": "2023-07-04T21:11:08Z", "updated_at": "2023-07-04T21:11:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2307.01923v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2307.01923", "identifiers": {"arxiv": "2307.01923"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2307.01923"}], "name": "An Algorithm for Persistent Homology Computation Using Homomorphic Encryption", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-07-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-07-04", "url": "https://arxiv.org/abs/2307.01923"}
{"aliases": [], "authors": ["Max Hort", "Anastasiia Grishina", "Leon Moonen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea2ae720a4cad7bdcb5f210fd82cb3bfe5a24de2ce370aea92650e87528f9559", "submitted_at": "2023-07-05T17:13:00Z", "updated_at": "2023-07-05T17:13:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2307.02443v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2307.02443", "identifiers": {"arxiv": "2307.02443"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2307.02443"}], "name": "An Exploratory Literature Study on Sharing and Energy Use of Language Models for Source Code", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-07-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-07-05", "url": "https://arxiv.org/abs/2307.02443"}
{"aliases": [], "authors": ["Wladimir Silva"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f54712365dee089c3869ee6b8d7cf9f4720aaebe895675b7cb807eef7bf16fac", "submitted_at": "2023-07-10T19:30:08Z", "updated_at": "2023-07-14T15:52:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2307.04867v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2307.04867", "identifiers": {"arxiv": "2307.04867"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2307.04867"}], "name": "Measurement Noise Mitigation in a Quantum Computer Using Image Intensity Filters", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-07-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-07-14", "url": "https://arxiv.org/abs/2307.04867"}
{"aliases": [], "authors": ["Chen Qian", "Wei Liu", "Hongzhang Liu", "Nuo Chen", "Yufan Dang", "Jiahao Li", "Cheng Yang", "Weize Chen", "Yusheng Su", "Xin Cong", "Juyuan Xu", "Dahai Li", "Zhiyuan Liu", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4aa017a83671f0339cbcbd29f6bb0eb006673116bacdfb92d287aa1e1c6c6920", "submitted_at": "2023-07-16T02:11:34Z", "updated_at": "2024-06-05T13:23:49Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2307.07924v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2307.07924", "identifiers": {"arxiv": "2307.07924"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2307.07924"}], "name": "ChatDev: Communicative Agents for Software Development", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-07-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-05", "url": "https://arxiv.org/abs/2307.07924"}
{"aliases": [], "authors": ["Christopher Rawles", "Alice Li", "Daniel Rodriguez", "Oriana Riva", "Timothy Lillicrap"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "def4616f7cb13fe839e500c4b3171eafbe37b155b5d14301137806d6a5cde939", "submitted_at": "2023-07-19T15:57:24Z", "updated_at": "2023-10-27T14:24:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2307.10088v2"}, "publication": {"authors": ["Christopher Rawles", "Alice Li", "Daniel Rodriguez", "Oriana Riva", "Timothy Lillicrap"], "date_basis": "Explicit publisher citation_publication_date: 2023-12-15. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2023, 12, 15], "doi": "10.52202/075280-2609", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "59708-59728", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/bbbb6308b402fe909c39dd29950c32e0-Paper-Datasets_and_Benchmarks.pdf", "publisher": null, "series": null, "title": "AndroidInTheWild: A Large-Scale Dataset For Android Device Control", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html", "venue": "Advances in Neural Information Processing Systems", "volume": "36"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "7fa75eae1a17878ef28414f64e1c46cc83539008f2644331b762248ec96efce5", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2307.10088", "identifiers": {"arxiv": "2307.10088"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2307.10088"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/bbbb6308b402fe909c39dd29950c32e0-Paper-Datasets_and_Benchmarks.pdf"}], "name": "Android in the Wild: A Large-Scale Dataset for Android Device Control", "notes": "Annotation reviewed against the primary abstract. Original dataset scale and experimental results are not independently reproduced.", "published_at": "2023-07-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Collects Android interaction demonstrations with screens, actions, and instructions across device and platform variations for robustness evaluation.", "topics": ["data-generation", "mobile-navigation", "evaluation"], "updated_at": "2023-10-27", "url": "https://arxiv.org/abs/2307.10088"}
{"aliases": [], "authors": ["Eugene Bagdasaryan", "Tsung-Yin Hsieh", "Ben Nassi", "Vitaly Shmatikov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b36ad344065d441084e608d42dd70c9f5ff9d777bb1ed8924188d3bda10f8373", "submitted_at": "2023-07-19T23:03:20Z", "updated_at": "2023-10-03T17:03:10Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2307.10490v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2307.10490", "identifiers": {"arxiv": "2307.10490"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2307.10490"}], "name": "Abusing Images and Sounds for Indirect Instruction Injection in Multi-Modal LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-07-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-03", "url": "https://arxiv.org/abs/2307.10490"}
{"aliases": [], "authors": ["Izzeddin Gur", "Hiroki Furuta", "Austin Huang", "Mustafa Safdari", "Yutaka Matsuo", "Douglas Eck", "Aleksandra Faust"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65328d525a6458ef8c401fb0ac4a9b6a5142bad462bd15a212d124c83a56209f", "submitted_at": "2023-07-24T14:56:30Z", "updated_at": "2024-02-25T16:17:43Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2307.12856v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2307.12856", "identifiers": {"arxiv": "2307.12856"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2307.12856"}], "name": "A Real-World WebAgent with Planning, Long Context Understanding, and Program Synthesis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-07-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines task decomposition, HTML summarization, and generated Python programs to carry out instructions on websites.", "topics": ["web-navigation", "planning"], "updated_at": "2024-02-25", "url": "https://arxiv.org/abs/2307.12856"}
{"aliases": [], "authors": ["Shuyan Zhou", "Frank F. Xu", "Hao Zhu", "Xuhui Zhou", "Robert Lo", "Abishek Sridhar", "Xianyi Cheng", "Tianyue Ou", "Yonatan Bisk", "Daniel Fried", "Uri Alon", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e9eca0165573ba53b663acbb1c7962fd2d64bc6633f10397f9cad229772a363", "submitted_at": "2023-07-25T22:59:32Z", "updated_at": "2024-04-16T15:13:18Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2307.13854v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2307.13854", "identifiers": {"arxiv": "2307.13854"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2307.13854"}], "name": "WebArena: A Realistic Web Environment for Building Autonomous Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2023-07-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces reproducible web environments with task-completion checks for evaluating autonomous web agents.", "topics": ["evaluation", "web-navigation", "long-horizon"], "updated_at": "2024-04-16", "url": "https://arxiv.org/abs/2307.13854"}
{"aliases": [], "authors": ["Erfan Shayegani", "Yue Dong", "Nael Abu-Ghazaleh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af5b26809fa8a1b0b0008599c6c28d4b14be51b86e093dae59540be1bd8c56a4", "submitted_at": "2023-07-26T23:11:15Z", "updated_at": "2023-10-10T22:17:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2307.14539v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2307.14539", "identifiers": {"arxiv": "2307.14539"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2307.14539"}], "name": "Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-07-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-10", "url": "https://arxiv.org/abs/2307.14539"}
{"aliases": [], "authors": ["Andy Zou", "Zifan Wang", "Nicholas Carlini", "Milad Nasr", "J. Zico Kolter", "Matt Fredrikson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "135d3086f2f6b6a9b097609a0c230d28f134dc96631c45965da87010fb6fe3e8", "submitted_at": "2023-07-27T17:49:12Z", "updated_at": "2023-12-20T20:48:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2307.15043v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2307.15043", "identifiers": {"arxiv": "2307.15043"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2307.15043"}], "name": "Universal and Transferable Adversarial Attacks on Aligned Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-07-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-20", "url": "https://arxiv.org/abs/2307.15043"}
{"aliases": [], "authors": ["Sirui Hong", "Mingchen Zhuge", "Jiaqi Chen", "Xiawu Zheng", "Yuheng Cheng", "Ceyao Zhang", "Jinlin Wang", "Zili Wang", "Steven Ka Shing Yau", "Zijuan Lin", "Liyang Zhou", "Chenyu Ran", "Lingfeng Xiao", "Chenglin Wu", "Jürgen Schmidhuber"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1209743a0f6299942572c774885dcbcdfa6f7ca5d68abf6f275bf4cd538f2a31", "submitted_at": "2023-08-01T07:49:10Z", "updated_at": "2024-11-01T14:36:52Z", "version": 7, "versioned_url": "https://arxiv.org/abs/2308.00352v7"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2308.00352", "identifiers": {"arxiv": "2308.00352"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.00352"}], "name": "MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-01", "url": "https://arxiv.org/abs/2308.00352"}
{"aliases": [], "authors": ["Po-Lin Chen", "Cheng-Shang Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7723a818fb4400422565835590b3b1489540b67b669d49fc678b474455520d2d", "submitted_at": "2023-08-03T06:19:58Z", "updated_at": "2023-08-03T06:19:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2308.01552v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2308.01552", "identifiers": {"arxiv": "2308.01552"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.01552"}], "name": "InterAct: Exploring the Potentials of ChatGPT as a Cooperative Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-03", "url": "https://arxiv.org/abs/2308.01552"}
{"aliases": [], "authors": ["Rodrigo Pedro", "Daniel Castro", "Paulo Carreira", "Nuno Santos"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c03a2f49769e42104ead027ef68972638189d643d36eb6d14276ef39cdf21f2", "submitted_at": "2023-08-03T19:03:18Z", "updated_at": "2025-01-27T20:37:59Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2308.01990v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "id": "arxiv:2308.01990", "identifiers": {"arxiv": "2308.01990"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.01990"}], "name": "From Prompt Injections to SQL Injection Attacks: How Protected is Your LLM-Integrated Web Application?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-27", "url": "https://arxiv.org/abs/2308.01990"}
{"aliases": [], "authors": ["Jingqing Ruan", "Yihong Chen", "Bin Zhang", "Zhiwei Xu", "Tianpeng Bao", "Guoqing Du", "Shiwei Shi", "Hangyu Mao", "Ziyue Li", "Xingyu Zeng", "Rui Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d7b54c9de66080089d25cf3cec6e732675c9ada01dafe7c52ec23a5730e3cb1", "submitted_at": "2023-08-07T09:22:03Z", "updated_at": "2025-12-29T14:46:42Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2308.03427v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2308.03427", "identifiers": {"arxiv": "2308.03427"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.03427"}], "name": "TPTU: Large Language Model-based AI Agents for Task Planning and Tool Usage", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-29", "url": "https://arxiv.org/abs/2308.03427"}
{"aliases": [], "authors": ["Xiao Liu", "Hao Yu", "Hanchen Zhang", "Yifan Xu", "Xuanyu Lei", "Hanyu Lai", "Yu Gu", "Hangliang Ding", "Kaiwen Men", "Kejuan Yang", "Shudan Zhang", "Xiang Deng", "Aohan Zeng", "Zhengxiao Du", "Chenhui Zhang", "Sheng Shen", "Tianjun Zhang", "Yu Su", "Huan Sun", "Minlie Huang", "Yuxiao Dong", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8c4ba4db4aeed3a7b7ce7f77d610d7e34b0e8db26a36463e2c0bdffcad8f22a", "submitted_at": "2023-08-07T16:08:11Z", "updated_at": "2025-10-04T03:54:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2308.03688v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2308.03688", "identifiers": {"arxiv": "2308.03688"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2308.03688"}], "name": "AgentBench: Evaluating LLMs as Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-08-07", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Evaluates language agents across eight interactive environments and analyzes failures in reasoning, decisions, and instruction following.", "topics": ["evaluation", "planning"], "updated_at": "2025-10-04", "url": "https://arxiv.org/abs/2308.03688"}
{"aliases": [], "authors": ["Brendan Pankovich", "Angus Kan", "Kwok Ho Wan", "Maike Ostmann", "Alex Neville", "Srikrishna Omkar", "Adel Sohbi", "Kamil Brádler"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e11a9d65f10fb8f5607d281f877e300a108f92e77fe1f5f5f69dbca59cb9d100", "submitted_at": "2023-08-08T11:36:23Z", "updated_at": "2023-08-08T11:36:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2308.04192v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2308.04192", "identifiers": {"arxiv": "2308.04192"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.04192"}], "name": "High photon-loss threshold quantum computing using GHZ-state measurements", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-08", "url": "https://arxiv.org/abs/2308.04192"}
{"aliases": [], "authors": ["Sivan Schwartz", "Avi Yaeli", "Segev Shlomov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08a301a8893c946657a5a3a10a395821a4908082bb78882edf39d57eac64dbf8", "submitted_at": "2023-08-10T07:12:11Z", "updated_at": "2023-08-10T07:12:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2308.05391v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2308.05391", "identifiers": {"arxiv": "2308.05391"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.05391"}], "name": "Enhancing Trust in LLM-Based AI Automation Agents: New Considerations and Future Challenges", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-10", "url": "https://arxiv.org/abs/2308.05391"}
{"aliases": [], "authors": ["Junxiao Shen", "John Dudley", "Per Ola Kristensson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40a7e2559e547fafa8bc5c1b641cd216b6330d491efd5de2e420d01e5d0d41ae", "submitted_at": "2023-08-10T18:43:44Z", "updated_at": "2024-10-18T07:24:54Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2308.05822v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2308.05822", "identifiers": {"arxiv": "2308.05822"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.05822"}], "name": "Encode-Store-Retrieve: Augmenting Human Memory through Language-Encoded Egocentric Perception", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-18", "url": "https://arxiv.org/abs/2308.05822"}
{"aliases": [], "authors": ["Yonatan Bitton", "Hritik Bansal", "Jack Hessel", "Rulin Shao", "Wanrong Zhu", "Anas Awadalla", "Josh Gardner", "Rohan Taori", "Ludwig Schmidt"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8a884cbb241f3768b88ba5b429bc2075120e8afe44319e456fcece6371c744e", "submitted_at": "2023-08-12T15:27:51Z", "updated_at": "2023-12-26T15:57:47Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2308.06595v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2308.06595", "identifiers": {"arxiv": "2308.06595"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.06595"}], "name": "VisIT-Bench: A Benchmark for Vision-Language Instruction Following Inspired by Real-World Use", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-26", "url": "https://arxiv.org/abs/2308.06595"}
{"aliases": [], "authors": ["Chi-Min Chan", "Weize Chen", "Yusheng Su", "Jianxuan Yu", "Wei Xue", "Shanghang Zhang", "Jie Fu", "Zhiyuan Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e284a82cda689e925fae293d2daf7f8faf14d7f3fa989b3478129ea3df7259ab", "submitted_at": "2023-08-14T15:13:04Z", "updated_at": "2023-08-14T15:13:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2308.07201v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2308.07201", "identifiers": {"arxiv": "2308.07201"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.07201"}], "name": "ChatEval: Towards Better LLM-based Evaluators through Multi-Agent Debate", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-14", "url": "https://arxiv.org/abs/2308.07201"}
{"aliases": [], "authors": ["Qingyun Wu", "Gagan Bansal", "Jieyu Zhang", "Yiran Wu", "Beibin Li", "Erkang Zhu", "Li Jiang", "Xiaoyun Zhang", "Shaokun Zhang", "Jiale Liu", "Ahmed Hassan Awadallah", "Ryen W White", "Doug Burger", "Chi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2504f7ea12083689aa6346511f6e3f52bab31f31f75b74e91e6c71f2b2a23038", "submitted_at": "2023-08-16T05:57:52Z", "updated_at": "2023-10-03T20:47:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2308.08155v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2308.08155", "identifiers": {"arxiv": "2308.08155"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.08155"}], "name": "AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-03", "url": "https://arxiv.org/abs/2308.08155"}
{"aliases": [], "authors": ["Sherif Abdelfattah", "Adrian Brown", "Pushi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5cfe80eeeef2378c771336fa1319127d01f63abe467073d4a0f921498a2b0005", "submitted_at": "2023-08-18T04:19:36Z", "updated_at": "2023-11-10T21:43:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2308.09289v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2308.09289", "identifiers": {"arxiv": "2308.09289"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.09289"}], "name": "Preference-conditioned Pixel-based AI Agent For Game Testing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-10", "url": "https://arxiv.org/abs/2308.09289"}
{"aliases": [], "authors": ["Jonas Kantic", "Fabian C. Legl", "Walter Stechele", "Jakob Hermann"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "515633a89f1cb6e8219133a38c6ce4b50663ae463fad073aa5e52bb06b17a4f8", "submitted_at": "2023-08-22T15:52:37Z", "updated_at": "2023-08-22T15:52:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2308.11522v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2308.11522", "identifiers": {"arxiv": "2308.11522"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.11522"}], "name": "ReLiCADA -- Reservoir Computing using Linear Cellular Automata Design Algorithm", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-22", "url": "https://arxiv.org/abs/2308.11522"}
{"aliases": [], "authors": ["Mahdi Manavi", "Yunpeng Zhang", "Guoning Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a29170e23a4f79b5e9c8194206d3350d5bfb4c64bf7b86c75b5710a9d4d5ce9", "submitted_at": "2023-08-22T20:41:13Z", "updated_at": "2023-08-22T20:41:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2308.11782v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2308.11782", "identifiers": {"arxiv": "2308.11782"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.11782"}], "name": "Resource Allocation in Cloud Computing Using Genetic Algorithm and Neural Network", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-22", "url": "https://arxiv.org/abs/2308.11782"}
{"aliases": [], "authors": ["Brandon Redding", "Joseph B. Murray", "Joseph D. Hart", "Zheyuan Zhu", "Shuo S. Pang", "Raktim Sarma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f217c430bac81b083d6485cc5922847ba9f0e04c53d4a449cedf5b49e50d052", "submitted_at": "2023-08-28T11:29:11Z", "updated_at": "2023-08-28T11:29:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2308.14504v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2308.14504", "identifiers": {"arxiv": "2308.14504"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.14504"}], "name": "Fiber optic computing using distributed feedback", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-08-28", "url": "https://arxiv.org/abs/2308.14504"}
{"aliases": [], "authors": ["Luiz O. R. Solak", "Daniel Z. Rossatto", "Celso J. Villas-Boas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ab5df86d5dfaf16c1b67939c07c068997269557cd1e49ddb1dcc803b300aae72", "submitted_at": "2023-08-28T20:09:54Z", "updated_at": "2024-06-27T20:27:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2308.14881v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2308.14881", "identifiers": {"arxiv": "2308.14881"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2308.14881"}], "name": "Universal quantum computation using atoms in cross-cavity systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-08-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-27", "url": "https://arxiv.org/abs/2308.14881"}
{"aliases": [], "authors": ["Hao Wen", "Yuanchun Li", "Guohong Liu", "Shanhui Zhao", "Tao Yu", "Toby Jia-Jun Li", "Shiqi Jiang", "Yunhao Liu", "Yaqin Zhang", "Yunxin Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a103fe10850a39d25bbb2b4a9ea0593d686938243bb912c674fd906e754c1fb", "submitted_at": "2023-08-29T13:02:30Z", "updated_at": "2024-03-09T09:38:51Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2308.15272v4"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2308.15272", "identifiers": {"arxiv": "2308.15272"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2308.15272"}], "name": "AutoDroid: LLM-powered Task Automation in Android", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-08-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines language-model reasoning with app analysis, UI representations, and exploration-derived memory for Android task automation.", "topics": ["mobile-navigation", "memory"], "updated_at": "2024-03-09", "url": "https://arxiv.org/abs/2308.15272"}
{"aliases": [], "authors": ["J. Hamilton", "K. Xu", "A. Brown", "C. A. Baron"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f5efe3b79e51f1fc96614f12984ac8d3641e98f869e149a894075a655d2f581", "submitted_at": "2023-09-05T15:36:24Z", "updated_at": "2023-09-05T15:36:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2309.02319v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2309.02319", "identifiers": {"arxiv": "2309.02319"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.02319"}], "name": "Robust frequency-dependent diffusion kurtosis computation using an efficient direction scheme, axisymmetric modelling, and spatial regularization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-09-05", "url": "https://arxiv.org/abs/2309.02319"}
{"aliases": [], "authors": ["Ozan Unal", "Christos Sakaridis", "Suman Saha", "Luc Van Gool"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7ff04bcf48f41378da8a8b55ed7700df7a67157a9d12189223bde7e94da9fb06", "submitted_at": "2023-09-08T19:27:01Z", "updated_at": "2024-07-16T08:14:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2309.04561v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2309.04561", "identifiers": {"arxiv": "2309.04561"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.04561"}], "name": "Four Ways to Improve Verbo-visual Fusion for Dense 3D Visual Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-16", "url": "https://arxiv.org/abs/2309.04561"}
{"aliases": [], "authors": ["Yiming Zhang", "ZeMing Gong", "Angel X. Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a3aa030d97353c0e0d77fa7956509b38316a9497250ecde52479d2c19893bde4", "submitted_at": "2023-09-11T06:03:39Z", "updated_at": "2023-09-11T06:03:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2309.05251v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2309.05251", "identifiers": {"arxiv": "2309.05251"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.05251"}], "name": "Multi3DRefer: Grounding Text Description to Multiple 3D Objects", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-09-11", "url": "https://arxiv.org/abs/2309.05251"}
{"aliases": [], "authors": ["Abigail McClain Gomez", "Taylor L. Patti", "Anima Anandkumar", "Susanne F. Yelin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "429f10be4783831cf87ccc3eb49e7bf8baedddfa16b74029cac6e62c6f88684f", "submitted_at": "2023-09-11T18:00:00Z", "updated_at": "2023-09-11T18:00:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2309.05693v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2309.05693", "identifiers": {"arxiv": "2309.05693"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.05693"}], "name": "Near-Term Distributed Quantum Computation using Mean-Field Corrections and Auxiliary Qubits", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-09-11", "url": "https://arxiv.org/abs/2309.05693"}
{"aliases": [], "authors": ["Pengzhou Cheng", "Zongru Wu", "Wei Du", "Haodong Zhao", "Wei Lu", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe4c2108f057be9130820b89f69f3621fa7ecd56ded0c980e015cd3478cd01f0", "submitted_at": "2023-09-12T08:48:38Z", "updated_at": "2024-12-04T05:04:46Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2309.06055v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2309.06055", "identifiers": {"arxiv": "2309.06055"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.06055"}], "name": "Backdoor Attacks and Countermeasures in Natural Language Processing Models: A Comprehensive Security Review", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-04", "url": "https://arxiv.org/abs/2309.06055"}
{"aliases": [], "authors": ["Jayjun Lee", "Adam J. Spiers"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6bf2204d7c138276a40c0b13f414401e371ed5d66f0957f2fac8998c118636d6", "submitted_at": "2023-09-14T09:26:03Z", "updated_at": "2023-09-14T09:26:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2309.07550v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2309.07550", "identifiers": {"arxiv": "2309.07550"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.07550"}], "name": "Naturalistic Robot Arm Trajectory Generation via Representation Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-09-14", "url": "https://arxiv.org/abs/2309.07550"}
{"aliases": [], "authors": ["Cameron Calcluth", "Nicolas Reichel", "Alessandro Ferraro", "Giulia Ferrini"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c4269deba97d892b29400d92784f8ec60be5ebe5155e82632f36be09c0cbe41", "submitted_at": "2023-09-14T16:15:14Z", "updated_at": "2024-01-05T11:08:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2309.07820v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2309.07820", "identifiers": {"arxiv": "2309.07820"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.07820"}], "name": "Sufficient condition for universal quantum computation using bosonic circuits", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-01-05", "url": "https://arxiv.org/abs/2309.07820"}
{"aliases": [], "authors": ["Kaixin Ma", "Hongming Zhang", "Hongwei Wang", "Xiaoman Pan", "Wenhao Yu", "Dong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "399c97382cac325a9d60a073a14221dee6ee6254cb0b8aa88a5341a7bb758c97", "submitted_at": "2023-09-15T05:44:08Z", "updated_at": "2024-02-21T17:42:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2309.08172v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2309.08172", "identifiers": {"arxiv": "2309.08172"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2309.08172"}], "name": "LASER: LLM Agent with State-Space Exploration for Web Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-09-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "LASER represents web navigation as exploration over a defined set of states, allowing an agent to revisit earlier states after mistakes. The study evaluates this backtracking approach on WebShop and Amazon tasks.", "topics": ["planning", "robustness", "web-navigation"], "updated_at": "2024-02-21", "url": "https://arxiv.org/abs/2309.08172"}
{"aliases": [], "authors": ["Yadong Lu", "Chunyuan Li", "Haotian Liu", "Jianwei Yang", "Jianfeng Gao", "Yelong Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "16f94a6b3b626befe340608e1292b3f52fbf4d98ea013a6c5acf8ec0426160ff", "submitted_at": "2023-09-18T17:30:46Z", "updated_at": "2023-09-18T17:30:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2309.09958v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2309.09958", "identifiers": {"arxiv": "2309.09958"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.09958"}], "name": "An Empirical Study of Scaling Instruct-Tuned Large Multimodal Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-09-18", "url": "https://arxiv.org/abs/2309.09958"}
{"aliases": [], "authors": ["Chenglong Wang", "John Thompson", "Bongshin Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f2ce9a39e61cd897a7731aadb8a9384bdb56740330f93557345eb320484fe814", "submitted_at": "2023-09-18T19:06:29Z", "updated_at": "2023-10-27T18:24:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2309.10094v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2309.10094", "identifiers": {"arxiv": "2309.10094"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.10094"}], "name": "Data Formulator: AI-powered Concept-driven Visualization Authoring", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-27", "url": "https://arxiv.org/abs/2309.10094"}
{"aliases": [], "authors": ["Jiahao Yu", "Xingwei Lin", "Zheng Yu", "Xinyu Xing"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "be7d747c17477d1026a6c723fedbb25f7c9351bf822da0d3a379154d54f64002", "submitted_at": "2023-09-19T02:19:48Z", "updated_at": "2024-06-27T16:01:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2309.10253v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2309.10253", "identifiers": {"arxiv": "2309.10253"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.10253"}], "name": "GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-27", "url": "https://arxiv.org/abs/2309.10253"}
{"aliases": [], "authors": ["Zhuosheng Zhang", "Aston Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d9cff1ece6115e87b7a9032f5720deb49694fbf570bea1f152018c4fa80d981", "submitted_at": "2023-09-20T16:12:32Z", "updated_at": "2024-06-07T04:52:29Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2309.11436v4"}, "publication": {"authors": ["Zhuosheng Zhang", "Aston Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.findings-acl.186", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-acl.186/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "3132-3149", "pdf_url": "https://aclanthology.org/2024.findings-acl.186.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "You Only Look at Screens: Multimodal Chain-of-Action Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-acl.186/", "venue": "Findings of the Association for Computational Linguistics: ACL 2024", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:48Z", "role": "primary", "sha256": "80fb0185db71f714655c22f5e1eb46d890c045b8af09c49d1b5ee2e0b6911478", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-acl.186/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2309.11436", "identifiers": {"arxiv": "2309.11436"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2309.11436"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-acl.186/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-acl.186.pdf"}], "name": "You Only Look at Screens: Multimodal Chain-of-Action Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-09-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Auto-GUI, a screenshot-based agent using previous actions and future action plans to guide device interaction.", "topics": ["mobile-navigation", "planning"], "updated_at": "2024-06-07", "url": "https://arxiv.org/abs/2309.11436"}
{"aliases": [], "authors": ["Tong Zhang", "X. Jessie Yang", "Boyang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ba19f6f7a95ed8d83468d4ba8c51427f358bef988d5107b572a36634691ce9a8", "submitted_at": "2023-09-25T09:00:38Z", "updated_at": "2024-08-12T08:34:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2309.13965v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2309.13965", "identifiers": {"arxiv": "2309.13965"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.13965"}], "name": "May I Ask a Follow-up Question? Understanding the Benefits of Conversations in Neural Network Explainability", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-12", "url": "https://arxiv.org/abs/2309.13965"}
{"aliases": [], "authors": ["Daniel Silver", "Tirthak Patel", "Devesh Tiwari"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4f55a7f2d4bdf2b0d15449ed63a91d116e649161942ef683509615b3df631f6", "submitted_at": "2023-09-26T16:36:11Z", "updated_at": "2023-09-26T16:36:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2309.15056v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2309.15056", "identifiers": {"arxiv": "2309.15056"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.15056"}], "name": "QUILT: Effective Multi-Class Classification on Quantum Computers Using an Ensemble of Diverse Quantum Classifiers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-09-26", "url": "https://arxiv.org/abs/2309.15056"}
{"aliases": [], "authors": ["Yangjun Ruan", "Honghua Dong", "Andrew Wang", "Silviu Pitis", "Yongchao Zhou", "Jimmy Ba", "Yann Dubois", "Chris J. Maddison", "Tatsunori Hashimoto"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45d2fa4ff64c91dcdc4fdb75380e72ad9fc56a892cbe84cc2bf3bc7099b0b494", "submitted_at": "2023-09-25T17:08:02Z", "updated_at": "2024-05-17T17:17:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2309.15817v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2309.15817", "identifiers": {"arxiv": "2309.15817"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2309.15817"}], "name": "Identifying the Risks of LM Agents with an LM-Emulated Sandbox", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-17", "url": "https://arxiv.org/abs/2309.15817"}
{"aliases": [], "authors": ["Yue Wu", "Xuan Tang", "Tom M. Mitchell", "Yuanzhi Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08255dcd9945c1b29810a87c23e2be4c3d69a12b63df3bcf1c9e08eae9cd7054", "submitted_at": "2023-10-02T18:52:11Z", "updated_at": "2024-03-17T23:23:31Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2310.01557v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2310.01557", "identifiers": {"arxiv": "2310.01557"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.01557"}], "name": "SmartPlay: A Benchmark for LLMs as Intelligent Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-17", "url": "https://arxiv.org/abs/2310.01557"}
{"aliases": [], "authors": ["Zheng-Xin Yong", "Cristina Menghini", "Stephen H. Bach"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bc4b70de889deddc2649a6dd9e21f9304b32fabfeb489b4370c218c1d91f32f5", "submitted_at": "2023-10-03T21:30:56Z", "updated_at": "2024-01-27T22:54:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.02446v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2310.02446", "identifiers": {"arxiv": "2310.02446"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.02446"}], "name": "Low-Resource Languages Jailbreak GPT-4", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-01-27", "url": "https://arxiv.org/abs/2310.02446"}
{"aliases": [], "authors": ["Paloma Sodhi", "S. R. K. Branavan", "Yoav Artzi", "Ryan McDonald"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d843f3c65e5961da773ec0265c2cfbf51d7c8214d9232ecad1638cc0dbb9ef64", "submitted_at": "2023-10-05T17:40:09Z", "updated_at": "2024-08-08T18:00:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2310.03720v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2310.03720", "identifiers": {"arxiv": "2310.03720"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.03720"}], "name": "SteP: Stacked LLM Policies for Web Actions", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-10-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "SteP dynamically composes specialized web-action policies through a stack that records the active chain of policy calls. It studies this control structure across WebArena, MiniWoB++, and CRM workflows.", "topics": ["planning", "web-navigation", "workflow"], "updated_at": "2024-08-08", "url": "https://arxiv.org/abs/2310.03720"}
{"aliases": [], "authors": ["Ash Jennings", "Xianjing Zhou", "Ivan Grytsenko", "Erika Kawakami"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c8a6868e58d7b4e3c754620f904f976f79c0b0a4360cb5e5457d9e0e69ed0ab", "submitted_at": "2023-10-06T09:38:37Z", "updated_at": "2023-12-28T07:23:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2310.04119v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2310.04119", "identifiers": {"arxiv": "2310.04119"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.04119"}], "name": "Quantum computing using floating electrons on cryogenic substrates: Potential And Challenges", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-28", "url": "https://arxiv.org/abs/2310.04119"}
{"aliases": [], "authors": ["Zhizheng Zhang", "Wenxuan Xie", "Xiaoyi Zhang", "Yan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f389de80bb86a759df2cfc7000188286efcd73c99a43ede376c46f190832ef80", "submitted_at": "2023-10-07T07:22:41Z", "updated_at": "2023-10-07T07:22:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.04716v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.04716", "identifiers": {"arxiv": "2310.04716"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.04716"}], "name": "Reinforced UI Instruction Grounding: Towards a Generic UI Task Automation API", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains a screenshot-based instruction-grounding model to predict target coordinates, using reinforcement learning to improve spatial decoding.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2023-10-07", "url": "https://arxiv.org/abs/2310.04716"}
{"aliases": [], "authors": ["Yue Jiang", "Eldon Schoop", "Amanda Swearngin", "Jeffrey Nichols"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a49e879891017071fce4df0ff4b07bd9e8b5a6e14385af565549aadbde1bc1cf", "submitted_at": "2023-10-07T16:32:34Z", "updated_at": "2023-10-07T16:32:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.04869v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.04869", "identifiers": {"arxiv": "2310.04869"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.04869"}], "name": "ILuvUI: Instruction-tuned LangUage-Vision modeling of UIs from Machine Conversations", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "ILuvUI generates 335,000 conversational training examples paired with UI screenshots, covering questions, descriptions, and plans. The resulting vision-language model is assessed on element detection, response quality, and examples of multi-step UI navigation.", "topics": ["data-generation", "screen-understanding", "grounding", "planning"], "updated_at": "2023-10-07", "url": "https://arxiv.org/abs/2310.04869"}
{"aliases": [], "authors": ["Jiabo Ye", "Anwen Hu", "Haiyang Xu", "Qinghao Ye", "Ming Yan", "Guohai Xu", "Chenliang Li", "Junfeng Tian", "Qi Qian", "Ji Zhang", "Qin Jin", "Liang He", "Xin Alex Lin", "Fei Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "39d5cf803efe3abaf70a9b78483bb0e750a8741adec1c4b7f3c05b9754d88560", "submitted_at": "2023-10-08T11:33:09Z", "updated_at": "2023-10-08T11:33:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.05126v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.05126", "identifiers": {"arxiv": "2310.05126"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.05126"}], "name": "UReader: Universal OCR-free Visually-situated Language Understanding with Multimodal Large Language Model", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The interface connection is static webpage-screenshot understanding; browser control is not claimed.", "published_at": "2023-10-08", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces UReader, which combines unified instruction tuning, auxiliary text-reading tasks and shape-adaptive image crops for OCR-free visual language understanding. Its evaluation includes webpage screenshots alongside documents, tables, charts and natural images.", "topics": ["screen-understanding", "efficiency"], "updated_at": "2023-10-08", "url": "https://arxiv.org/abs/2310.05126"}
{"aliases": [], "authors": ["Zeming Wei", "Yifei Wang", "Ang Li", "Yichuan Mo", "Yisen Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e182f1c29bc882be1af5f3e5968ba2298184867c3210c77c4d210171968d68bd", "submitted_at": "2023-10-10T07:50:29Z", "updated_at": "2024-05-25T07:01:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2310.06387v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2310.06387", "identifiers": {"arxiv": "2310.06387"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.06387"}], "name": "Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-25", "url": "https://arxiv.org/abs/2310.06387"}
{"aliases": [], "authors": ["Carlos E. Jimenez", "John Yang", "Alexander Wettig", "Shunyu Yao", "Kexin Pei", "Ofir Press", "Karthik Narasimhan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67a7b5b540a58a91f9b74da32682d5cc16f6c3224401294320b024b317d98722", "submitted_at": "2023-10-10T16:47:29Z", "updated_at": "2024-11-11T23:05:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2310.06770v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.06770", "identifiers": {"arxiv": "2310.06770"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.06770"}], "name": "SWE-bench: Can Language Models Resolve Real-World GitHub Issues?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-11", "url": "https://arxiv.org/abs/2310.06770"}
{"aliases": [], "authors": ["Yuyang Liu", "Weijun Dong", "Yingdong Hu", "Chuan Wen", "Zhao-Heng Yin", "Chongjie Zhang", "Yang Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8789e505ec521b97cf172f81589c55be3175917a7a66a48fffa93430df3736ce", "submitted_at": "2023-10-11T12:34:39Z", "updated_at": "2024-02-07T14:43:41Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2310.07433v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.07433", "identifiers": {"arxiv": "2310.07433"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.07433"}], "name": "Imitation Learning from Observation with Automatic Discount Scheduling", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-07", "url": "https://arxiv.org/abs/2310.07433"}
{"aliases": [], "authors": ["Sumedh A Sontakke", "Jesse Zhang", "Sébastien M. R. Arnold", "Karl Pertsch", "Erdem Bıyık", "Dorsa Sadigh", "Chelsea Finn", "Laurent Itti"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e0cb3cb070ea95a702a9b2e20d3dba69723fc5259b371f2898d015c955baec8a", "submitted_at": "2023-10-11T21:10:21Z", "updated_at": "2023-10-11T21:10:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.07899v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.07899", "identifiers": {"arxiv": "2310.07899"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.07899"}], "name": "RoboCLIP: One Demonstration is Enough to Learn Robot Policies", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-11", "url": "https://arxiv.org/abs/2310.07899"}
{"aliases": [], "authors": ["Shaofei Cai", "Bowei Zhang", "Zihao Wang", "Xiaojian Ma", "Anji Liu", "Yitao Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9f1951c199e82856f6996774cf639f2504c8e903872f3f1f1e4f44942ff61fb1", "submitted_at": "2023-10-12T11:31:01Z", "updated_at": "2023-11-29T01:07:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.08235v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2310.08235", "identifiers": {"arxiv": "2310.08235"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.08235"}], "name": "GROOT: Learning to Follow Instructions by Watching Gameplay Videos", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-29", "url": "https://arxiv.org/abs/2310.08235"}
{"aliases": [], "authors": ["Xinyue Zheng", "Haowei Lin", "Kaichen He", "Zihao Wang", "Zilong Zheng", "Yitao Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "62bad2be8b6f6e7ed7d31f7bd1465d10b8861da3bcb44302ebac90b68a974a91", "submitted_at": "2023-10-12T14:38:25Z", "updated_at": "2025-06-03T02:30:05Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2310.08367v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2310.08367", "identifiers": {"arxiv": "2310.08367"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.08367"}], "name": "MCU: An Evaluation Framework for Open-Ended Game Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-03", "url": "https://arxiv.org/abs/2310.08367"}
{"aliases": [], "authors": ["Jingkang Yang", "Yuhao Dong", "Shuai Liu", "Bo Li", "Ziyue Wang", "Chencheng Jiang", "Haoran Tan", "Jiamu Kang", "Yuanhan Zhang", "Kaiyang Zhou", "Ziwei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af5f7992246baa7b724a7d30fcbedfd2d0823d47c1f5af0a71fcbe4c1782e8e0", "submitted_at": "2023-10-12T17:59:58Z", "updated_at": "2024-10-20T17:57:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.08588v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2310.08588", "identifiers": {"arxiv": "2310.08588"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.08588"}], "name": "Octopus: Embodied Vision-Language Programmer from Environmental Feedback", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-20", "url": "https://arxiv.org/abs/2310.08588"}
{"aliases": [], "authors": ["Tao Li", "Gang Li", "Zhiwei Deng", "Bryan Wang", "Yang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95092986184a232af62c61fa2aedefb837851dce9e0ff7a1128a67d47edfde40", "submitted_at": "2023-10-12T21:53:37Z", "updated_at": "2023-10-23T17:39:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2310.08740v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.08740", "identifiers": {"arxiv": "2310.08740"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.08740"}], "name": "A Zero-Shot Language Agent for Computer Control with Structured Reflection", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-10-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies a zero-shot computer-control agent that plans executable actions and organizes its own reflections after mistakes. MiniWoB++ experiments examine learning to complete tasks without supplied expert trajectories.", "topics": ["self-reflection", "planning", "web-navigation"], "updated_at": "2023-10-23", "url": "https://arxiv.org/abs/2310.08740"}
{"aliases": [], "authors": ["Han Wu", "Guangwu Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7d19c533cb05162a4ec25954e0b8b8a9c9f48adce88531fb95dd4e4f3878ff4", "submitted_at": "2023-10-13T13:49:05Z", "updated_at": "2023-11-07T08:29:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.09103v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.09103", "identifiers": {"arxiv": "2310.09103"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.09103"}], "name": "Qin's Algorithm, Continued Fractions and 2-dimensional Lattices", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-07", "url": "https://arxiv.org/abs/2310.09103"}
{"aliases": [], "authors": ["Shuyu Jiang", "Xingshu Chen", "Rui Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b50576886eeda57665aa7e3abd98975955fc0d820d478a0bd6e3e19486e70f16", "submitted_at": "2023-10-16T05:19:25Z", "updated_at": "2023-10-16T05:19:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.10077v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2310.10077", "identifiers": {"arxiv": "2310.10077"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.10077"}], "name": "Prompt Packer: Deceiving LLMs through Compositional Instruction with Hidden Attacks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-16", "url": "https://arxiv.org/abs/2310.10077"}
{"aliases": [], "authors": ["Tianbao Xie", "Fan Zhou", "Zhoujun Cheng", "Peng Shi", "Luoxuan Weng", "Yitao Liu", "Toh Jing Hua", "Junning Zhao", "Qian Liu", "Che Liu", "Leo Z. Liu", "Yiheng Xu", "Hongjin Su", "Dongchan Shin", "Caiming Xiong", "Tao Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15d7945fcf1ee4290761d34bbeef9e24bed82bb748bdb7aaa791457c42892fec", "submitted_at": "2023-10-16T17:54:53Z", "updated_at": "2023-10-16T17:54:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.10634v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:11:36Z", "role": "primary", "sha256": "90df698b27a336e2fb7f0ac60a3b8d3e3897c00035ab7968c593a5ab539d177a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.10634", "identifiers": {"arxiv": "2310.10634"}, "kind": "paper", "last_verified_at": "2026-10-17T17:11:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.10634"}], "name": "OpenAgents: An Open Platform for Language Agents in the Wild", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. Scope is adjacent: autonomous web browsing is one of three hosted agents, beside data-analysis and API-plugin agents.", "published_at": "2023-10-16", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "OpenAgents hosts three language agents, for data analysis, API plugins and autonomous web browsing, behind a web interface for non-expert users, with local deployment for developers.", "topics": ["web-navigation", "tool-integration"], "updated_at": "2023-10-16", "url": "https://arxiv.org/abs/2310.10634"}
{"aliases": [], "authors": ["Michele Coscia", "Karel Devriendt"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "83dacf536e94f372132419c0b1225cb0e6e52690f12d796cbc30009c48b5e90e", "submitted_at": "2023-10-17T12:55:14Z", "updated_at": "2023-10-17T12:55:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.11222v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2310.11222", "identifiers": {"arxiv": "2310.11222"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.11222"}], "name": "Fast Node Vector Distance Computations using Laplacian Solvers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-17", "url": "https://arxiv.org/abs/2310.11222"}
{"aliases": [], "authors": ["Jianwei Yang", "Hao Zhang", "Feng Li", "Xueyan Zou", "Chunyuan Li", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "abafe2b974ddd80416b215cd4c48d801e18fe9162b9c097d819d42eaff68a069", "submitted_at": "2023-10-17T17:51:31Z", "updated_at": "2023-11-06T07:39:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.11441v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.11441", "identifiers": {"arxiv": "2310.11441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.11441"}], "name": "Set-of-Mark Prompting Unleashes Extraordinary Visual Grounding in GPT-4V", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-10-17", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Overlays labeled image regions to help multimodal models refer to and reason about visual targets; a general visual-prompting method relevant to GUI grounding.", "topics": ["grounding"], "updated_at": "2023-11-06", "url": "https://arxiv.org/abs/2310.11441"}
{"aliases": [], "authors": ["Yuxin Wang", "Xueze Kang", "Shaohuai Shi", "Xin He", "Zhenheng Tang", "Xinglin Pan", "Yang Zheng", "Xiaoyu Wu", "Amelie Chi Zhou", "Bingsheng He", "Xiaowen Chu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c47a0486ab46dcec9cca79dcb9d8ee6430db1d2cedf2f81910816c58587122ae", "submitted_at": "2023-10-19T11:59:01Z", "updated_at": "2024-08-19T16:17:06Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2310.12670v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.12670", "identifiers": {"arxiv": "2310.12670"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.12670"}], "name": "Fault-Tolerant Hybrid-Parallel Training at Scale with Reliable and Efficient In-memory Checkpointing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-19", "url": "https://arxiv.org/abs/2310.12670"}
{"aliases": [], "authors": ["Lin Li", "Yifei Wang", "Chawin Sitawarin", "Michael Spratling"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc49c8a21781108d18f3e0b55d93854d99e9aa8787a6bfa0c2102efa3b21ad78", "submitted_at": "2023-10-19T14:50:46Z", "updated_at": "2024-06-04T03:01:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.12793v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.12793", "identifiers": {"arxiv": "2310.12793"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.12793"}], "name": "OODRobustBench: a Benchmark and Large-Scale Analysis of Adversarial Robustness under Distribution Shift", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-04", "url": "https://arxiv.org/abs/2310.12793"}
{"aliases": [], "authors": ["Yupei Liu", "Yuqi Jia", "Runpeng Geng", "Jinyuan Jia", "Neil Zhenqiang Gong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "74fc38f55fc1c9709028c61d9a9cacbbbc8f9a2dbc7400032582b61885b5957c", "submitted_at": "2023-10-19T15:12:09Z", "updated_at": "2025-11-12T08:51:29Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2310.12815v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.12815", "identifiers": {"arxiv": "2310.12815"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.12815"}], "name": "Formalizing and Benchmarking Prompt Injection Attacks and Defenses", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-12", "url": "https://arxiv.org/abs/2310.12815"}
{"aliases": [], "authors": ["Chunyi Sun", "Junlin Han", "Weijian Deng", "Xinlong Wang", "Zishan Qin", "Stephen Gould"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "55c387550128d3971c857450cbec73f43e390faf49489359c006b8d459c479c5", "submitted_at": "2023-10-19T17:41:48Z", "updated_at": "2024-05-29T01:35:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.12945v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.12945", "identifiers": {"arxiv": "2310.12945"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.12945"}], "name": "3D-GPT: Procedural 3D Modeling with Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-29", "url": "https://arxiv.org/abs/2310.12945"}
{"aliases": [], "authors": ["Liyiming Ke", "Yunchu Zhang", "Abhay Deshpande", "Siddhartha Srinivasa", "Abhishek Gupta"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb04fbd306e66e11c498d17c06cadd2a48840bf1b8c7dea65c8fe1dc6ae805d5", "submitted_at": "2023-10-19T17:59:03Z", "updated_at": "2024-06-03T21:37:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.12972v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.12972", "identifiers": {"arxiv": "2310.12972"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.12972"}], "name": "CCIL: Continuity-based Data Augmentation for Corrective Imitation Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-03", "url": "https://arxiv.org/abs/2310.12972"}
{"aliases": [], "authors": ["Xilie Xu", "Keyi Kong", "Ning Liu", "Lizhen Cui", "Di Wang", "Jingfeng Zhang", "Mohan Kankanhalli"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0575dc1e01cbad3f6321dd0ecfc995b7a8f0ad00d5d9e43163ee80a39170da8", "submitted_at": "2023-10-20T08:16:46Z", "updated_at": "2023-10-20T08:16:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.13345v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.13345", "identifiers": {"arxiv": "2310.13345"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.13345"}], "name": "An LLM can Fool Itself: A Prompt-Based Adversarial Attack", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-20", "url": "https://arxiv.org/abs/2310.13345"}
{"aliases": [], "authors": ["Justin Payan", "Swaroop Mishra", "Mukul Singh", "Carina Negreanu", "Christian Poelitz", "Chitta Baral", "Subhro Roy", "Rasika Chakravarthy", "Benjamin Van Durme", "Elnaz Nouri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1648acae2228b8ae2f042d7986585a66bc6a649320d569b12fd093b96c15cb5c", "submitted_at": "2023-10-23T02:00:55Z", "updated_at": "2023-10-23T02:00:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.14495v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:47Z", "role": "primary", "sha256": "40fd8a31083963f1cbbcfeb308244494139277e4a6d0e56c2ebe5a7cda821e7c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2310.14495v1"}], "id": "arxiv:2310.14495", "identifiers": {"arxiv": "2310.14495"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.14495"}], "name": "InstructExcel: A Benchmark for Natural Language Instruction in Excel", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced. This is an application-specific automation benchmark grounded in recorded Excel user actions. Its execution interface is OfficeScripts, so it is adjacent rather than a visual GUI navigation benchmark.", "published_at": "2023-10-23", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces InstructExcel, a benchmark for generating Excel OfficeScripts from natural-language requests, built using scripts recorded from user actions and a collection of existing spreadsheets.", "topics": ["evaluation", "workflow", "automation"], "updated_at": "2023-10-23", "url": "https://arxiv.org/abs/2310.14495"}
{"aliases": [], "authors": ["Heyi Tao", "Sethuraman T", "Michal Shlapentokh-Rothman", "Derek Hoiem"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7defced3e3beb77a5a6513dbfdff3f102edd922907b523bd062522e54720fa5e", "submitted_at": "2023-10-24T17:57:03Z", "updated_at": "2023-10-25T03:54:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.16042v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.16042", "identifiers": {"arxiv": "2310.16042"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.16042"}], "name": "WebWISE: Web Interface Control and Sequential Exploration with Large Language Models", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-10-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebWISE filters DOM elements into observations and generates short programs for click, scroll, and text-entry steps. It evaluates in-context learning from either one supplied example or an example collected during a successful zero-shot MiniWoB++ trial.", "topics": ["web-navigation", "tool-integration", "planning"], "updated_at": "2023-10-25", "url": "https://arxiv.org/abs/2310.16042"}
{"aliases": [], "authors": ["Nicholas X. Armendarez", "Ahmed S. Mohamed", "Anurag Dhungel", "Md Razuan Hossain", "Md Sakib Hasan", "Joseph S. Najem"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bceb700b0b286bdeace9fb35e5023aefe0d1a28ecfa06f6a42a78c03b81fb670", "submitted_at": "2023-10-25T03:27:43Z", "updated_at": "2023-10-25T03:27:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.16331v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2310.16331", "identifiers": {"arxiv": "2310.16331"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.16331"}], "name": "Brain-Inspired Reservoir Computing Using Memristors with Tunable Dynamics and Short-Term Plasticity", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-25", "url": "https://arxiv.org/abs/2310.16331"}
{"aliases": [], "authors": ["Bernard J. Giron Castro", "Christophe Peucheret", "Darko Zibar", "Francesco Da Ros"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4414c8aa9183d05f24b76972f40b6d5fa3257b84ff45b7b497c5c7947d2a1590", "submitted_at": "2023-10-25T12:24:56Z", "updated_at": "2024-04-27T18:25:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.16588v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2310.16588", "identifiers": {"arxiv": "2310.16588"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.16588"}], "name": "Multi-Task Wavelength-Multiplexed Reservoir Computing Using a Silicon Microring Resonator", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-27", "url": "https://arxiv.org/abs/2310.16588"}
{"aliases": [], "authors": ["Ajay Mandlekar", "Soroush Nasiriany", "Bowen Wen", "Iretiayo Akinola", "Yashraj Narang", "Linxi Fan", "Yuke Zhu", "Dieter Fox"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b7cc78bf57f3b0210db1cfb96fc7477c008697dc361ca9aa0dcc0093e2164b54", "submitted_at": "2023-10-26T17:17:31Z", "updated_at": "2023-10-26T17:17:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.17596v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.17596", "identifiers": {"arxiv": "2310.17596"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.17596"}], "name": "MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-10-26", "url": "https://arxiv.org/abs/2310.17596"}
{"aliases": [], "authors": ["Jiarun Liu", "Wentao Hu", "Chunhong Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c72eab40cdc28981cf83a753d88f5181b5441d82cd568265457fef96d75b36b6", "submitted_at": "2023-10-20T11:10:14Z", "updated_at": "2023-10-31T06:25:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.18331v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2310.18331", "identifiers": {"arxiv": "2310.18331"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.18331"}], "name": "AllTogether: Investigating the Efficacy of Spliced Prompt for Web Navigation using Large Language Models", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-10-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "AllTogether standardizes prompts for HTML-based web navigation and compares prompting with instruction fine-tuning. Its analysis examines how HTML context, interaction history, and current environmental feedback affect task execution.", "topics": ["web-navigation", "interaction-history", "evaluation"], "updated_at": "2023-10-31", "url": "https://arxiv.org/abs/2310.18331"}
{"aliases": [], "authors": ["Ziqi Zhang", "Xiao Xiong", "Zifeng Zhuang", "Jinxin Liu", "Donglin Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4d7f807c9709dabd17873c2ced6dacac16b7dd4377fd19e3ac8132ca23f7c869", "submitted_at": "2023-10-07T00:02:05Z", "updated_at": "2024-05-28T07:57:57Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2310.19805v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2310.19805", "identifiers": {"arxiv": "2310.19805"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2310.19805"}], "name": "Improving Offline-to-Online Reinforcement Learning with Q Conditioned State Entropy Exploration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-10-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-28", "url": "https://arxiv.org/abs/2310.19805"}
{"aliases": [], "authors": ["Sam Toyer", "Olivia Watkins", "Ethan Adrian Mendes", "Justin Svegliato", "Luke Bailey", "Tiffany Wang", "Isaac Ong", "Karim Elmaaroufi", "Pieter Abbeel", "Trevor Darrell", "Alan Ritter", "Stuart Russell"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "288e075e77e07be8745c31e41cead5207f956dec4bcf80665acc6152ae82939e", "submitted_at": "2023-11-02T06:13:36Z", "updated_at": "2023-11-02T06:13:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.01011v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2311.01011", "identifiers": {"arxiv": "2311.01011"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.01011"}], "name": "Tensor Trust: Interpretable Prompt Injection Attacks from an Online Game", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-02", "url": "https://arxiv.org/abs/2311.01011"}
{"aliases": [], "authors": ["David Schwerdt", "Lee Peleg", "Yotam Shapira", "Nadav Priel", "Yanay Florshaim", "Avram Gross", "Ayelet Zalic", "Gadi Afek", "Nitzan Akerman", "Ady Stern", "Amit Ben Kish", "Roee Ozeri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a62d6e7b8b614a319cdcf60d2a960196abcd881cf1fd3a32b85b7f4fe183bf93", "submitted_at": "2023-11-02T12:06:49Z", "updated_at": "2024-11-04T13:35:50Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2311.01168v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2311.01168", "identifiers": {"arxiv": "2311.01168"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.01168"}], "name": "Scalable architecture for trapped-ion quantum computing using RF traps and dynamic optical potentials", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-04", "url": "https://arxiv.org/abs/2311.01168"}
{"aliases": [], "authors": ["Jim Furches", "Sarah Chehade", "Kathleen Hamilton", "Nathan Wiebe", "Carlos Ortiz Marrero"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f455da827db0194e05ca06686e3d05239ebf0475b15fefde1fcf1fb51e01ad06", "submitted_at": "2023-11-02T16:17:18Z", "updated_at": "2025-06-28T07:42:43Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2311.01363v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2311.01363", "identifiers": {"arxiv": "2311.01363"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.01363"}], "name": "Application-level Benchmarking of Quantum Computers using Nonlocal Game Strategies", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-28", "url": "https://arxiv.org/abs/2311.01363"}
{"aliases": [], "authors": ["Yiduo Guo", "Zekai Zhang", "Yaobo Liang", "Dongyan Zhao", "Nan Duan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "58dfa75fd82a2cc1d41954f10feb1d987de6469782d11d04c3b8bfffc8a3b745", "submitted_at": "2023-11-03T08:06:35Z", "updated_at": "2023-11-07T10:13:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.01767v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2311.01767", "identifiers": {"arxiv": "2311.01767"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2311.01767"}], "name": "PPTC Benchmark: Evaluating Large Language Models for PowerPoint Task Completion", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-11-03", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Evaluates multi-turn PowerPoint creation and editing, scoring the resulting presentation files rather than requiring a fixed API sequence.", "topics": ["evaluation", "tool-integration"], "updated_at": "2023-11-07", "url": "https://arxiv.org/abs/2311.01767"}
{"aliases": [], "authors": ["Weihan Wang", "Qingsong Lv", "Wenmeng Yu", "Wenyi Hong", "Ji Qi", "Yan Wang", "Junhui Ji", "Zhuoyi Yang", "Lei Zhao", "Xixuan Song", "Jiazheng Xu", "Bin Xu", "Juanzi Li", "Yuxiao Dong", "Ming Ding", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6d81bc0a2838429fefe555698068c411db3e2865b81fb1a942d1eb919e0a051", "submitted_at": "2023-11-06T13:04:39Z", "updated_at": "2024-02-04T08:23:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.03079v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2311.03079", "identifiers": {"arxiv": "2311.03079"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.03079"}], "name": "CogVLM: Visual Expert for Pretrained Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-04", "url": "https://arxiv.org/abs/2311.03079"}
{"aliases": [], "authors": ["Edgar Andres Ruiz Guzman", "Denis Lacroix"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53121a28239d2ce78a1757d79a717b391805f6f6145b2649a5f2dba896214fde", "submitted_at": "2023-11-08T10:11:01Z", "updated_at": "2023-11-08T10:11:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.04571v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2311.04571", "identifiers": {"arxiv": "2311.04571"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.04571"}], "name": "Restoring symmetries in quantum computing using Classical Shadows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-08", "url": "https://arxiv.org/abs/2311.04571"}
{"aliases": [], "authors": ["Conor Mc Keever", "Michael Lubasch"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "85eed520299cd238f99d7927ef552d37c211e5db7f005f3d274f179c1a31f3ed", "submitted_at": "2023-11-09T17:45:07Z", "updated_at": "2024-06-20T15:53:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2311.05544v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2311.05544", "identifiers": {"arxiv": "2311.05544"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.05544"}], "name": "Towards adiabatic quantum computing using compressed quantum circuits", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-20", "url": "https://arxiv.org/abs/2311.05544"}
{"aliases": [], "authors": ["Zihao Wang", "Shaofei Cai", "Anji Liu", "Yonggang Jin", "Jinbing Hou", "Bowei Zhang", "Haowei Lin", "Zhaofeng He", "Zilong Zheng", "Yaodong Yang", "Xiaojian Ma", "Yitao Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2895d2132d51a4339e84cdf096f2189bc3ef00dbf42c46cbc12b7624282321b8", "submitted_at": "2023-11-10T11:17:58Z", "updated_at": "2023-11-30T07:39:48Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2311.05997v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2311.05997", "identifiers": {"arxiv": "2311.05997"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.05997"}], "name": "JARVIS-1: Open-World Multi-task Agents with Memory-Augmented Multimodal Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-30", "url": "https://arxiv.org/abs/2311.05997"}
{"aliases": [], "authors": ["Seungjun Moon", "Hyungjoo Chae", "Yongho Song", "Taeyoon Kwon", "Dongjin Kang", "Kai Tzu-iunn Ong", "Seung-won Hwang", "Jinyoung Yeo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "13a594e5264180b764af81d80e1c13e69fd0df84a04b2d4818aca5a7df6c81b9", "submitted_at": "2023-11-13T10:15:19Z", "updated_at": "2024-02-23T06:56:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2311.07215v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2311.07215", "identifiers": {"arxiv": "2311.07215"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.07215"}], "name": "Coffee: Boost Your Code LLMs by Fixing Bugs with Feedback", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-23", "url": "https://arxiv.org/abs/2311.07215"}
{"aliases": [], "authors": ["An Yan", "Zhengyuan Yang", "Wanrong Zhu", "Kevin Lin", "Linjie Li", "Jianfeng Wang", "Jianwei Yang", "Yiwu Zhong", "Julian McAuley", "Jianfeng Gao", "Zicheng Liu", "Lijuan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebe6ff09f876b81d033a064301a1c23e3c141d2a1ad0d7a5df91e729377fcd3b", "submitted_at": "2023-11-13T18:53:37Z", "updated_at": "2023-11-13T18:53:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.07562v1"}}, "environments": ["ios", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2311.07562", "identifiers": {"arxiv": "2311.07562"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2311.07562"}], "name": "GPT-4V in Wonderland: Large Multimodal Models for Zero-Shot Smartphone GUI Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-11-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies MM-Navigator, a multimodal agent that interprets smartphone screenshots and predicts localized actions without task-specific training.", "topics": ["mobile-navigation", "grounding"], "updated_at": "2023-11-13", "url": "https://arxiv.org/abs/2311.07562"}
{"aliases": [], "authors": ["Elizabeth Seger", "Noemi Dreksler", "Richard Moulange", "Emily Dardaman", "Jonas Schuett", "K. Wei", "Christoph Winter", "Mackenzie Arnold", "Seán Ó hÉigeartaigh", "Anton Korinek", "Markus Anderljung", "Ben Bucknall", "Alan Chan", "Eoghan Stafford", "Leonie Koessler", "Aviv Ovadya", "Ben Garfinkel", "Emma Bluemke", "Michael Aird", "Patrick Levermore", "Julian Hazell", "Abhishek Gupta"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "56f2a50b9bcf7969405d685352e46dda26fe8cc4c63715f83d01f479e068616d", "submitted_at": "2023-09-29T17:03:45Z", "updated_at": "2023-09-29T17:03:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.09227v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2311.09227", "identifiers": {"arxiv": "2311.09227"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.09227"}], "name": "Open-Sourcing Highly Capable Foundation Models: An evaluation of risks, benefits, and alternative methods for pursuing open-source objectives", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-09-29", "url": "https://arxiv.org/abs/2311.09227"}
{"aliases": [], "authors": ["Fangzhi Xu", "Zhiyong Wu", "Qiushi Sun", "Siyu Ren", "Fei Yuan", "Shuai Yuan", "Qika Lin", "Yu Qiao", "Jun Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8b60e5ace8c14ff850fef7a35225d7bfa6045922adfda3d6edd155ecea4830c", "submitted_at": "2023-11-15T18:59:56Z", "updated_at": "2024-02-18T06:24:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.09278v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2311.09278", "identifiers": {"arxiv": "2311.09278"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.09278"}], "name": "Symbol-LLM: Towards Foundational Symbol-centric Interface For Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-18", "url": "https://arxiv.org/abs/2311.09278"}
{"aliases": [], "authors": ["Lingbo Mo", "Boshi Wang", "Muhao Chen", "Huan Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b395a97cc604cad3ded2a7c6d7213ed1ccf7ba0e18da68c0cdd1ac9f22b614f3", "submitted_at": "2023-11-15T23:33:07Z", "updated_at": "2024-04-02T15:21:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.09447v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2311.09447", "identifiers": {"arxiv": "2311.09447"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.09447"}], "name": "How Trustworthy are Open-Source LLMs? An Assessment under Malicious Demonstrations Shows their Vulnerabilities", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-02", "url": "https://arxiv.org/abs/2311.09447"}
{"aliases": [], "authors": ["Adhitha Dias", "Logan Anderson", "Kirshanthan Sundararajah", "Artem Pelenitsyn", "Milind Kulkarni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e7b085a57fb509a684583a8f5830e83ff735567b8049fda1eafcb4a86e3a02f", "submitted_at": "2023-11-16T04:05:50Z", "updated_at": "2024-08-19T14:13:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2311.09549v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2311.09549", "identifiers": {"arxiv": "2311.09549"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.09549"}], "name": "SparseAuto: An Auto-Scheduler for Sparse Tensor Computations Using Recursive Loop Nest Restructuring", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-19", "url": "https://arxiv.org/abs/2311.09549"}
{"aliases": [], "authors": ["Silen Naihin", "David Atkinson", "Marc Green", "Merwane Hamadi", "Craig Swift", "Douglas Schonholtz", "Adam Tauman Kalai", "David Bau"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cbce7e5d2ad2b6d29136b52b0532eb118f71a926ddcac32ba672ce177c3132e6", "submitted_at": "2023-11-17T14:06:05Z", "updated_at": "2023-12-03T13:18:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2311.10538v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2311.10538", "identifiers": {"arxiv": "2311.10538"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.10538"}], "name": "Testing Language Model Agents Safely in the Wild", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-03", "url": "https://arxiv.org/abs/2311.10538"}
{"aliases": [], "authors": ["Yining Ye", "Xin Cong", "Shizuo Tian", "Jiannan Cao", "Hao Wang", "Yujia Qin", "Yaxi Lu", "Heyang Yu", "Huadong Wang", "Yankai Lin", "Zhiyuan Liu", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1cb7c2d81302a2dfa8179c5f83e940f5b2015c591dc779c770a8382694b67801", "submitted_at": "2023-11-02T14:32:16Z", "updated_at": "2023-11-23T12:14:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.10751v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2311.10751", "identifiers": {"arxiv": "2311.10751"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.10751"}], "name": "ProAgent: From Robotic Process Automation to Agentic Process Automation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-23", "url": "https://arxiv.org/abs/2311.10751"}
{"aliases": [], "authors": ["Jiaming Zhang", "Xingjun Ma", "Xin Wang", "Lingyu Qiu", "Jiaqi Wang", "Yu-Gang Jiang", "Jitao Sang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8021721830bae0c9ff15a1d70ae37d975faa04f400dd5ad976d9c1549db0d467", "submitted_at": "2023-11-19T07:47:43Z", "updated_at": "2024-08-19T11:38:47Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2311.11261v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2311.11261", "identifiers": {"arxiv": "2311.11261"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.11261"}], "name": "Adversarial Prompt Tuning for Vision-Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-19", "url": "https://arxiv.org/abs/2311.11261"}
{"aliases": [], "authors": ["Yilun Kong", "Jingqing Ruan", "Yihong Chen", "Bin Zhang", "Tianpeng Bao", "Shiwei Shi", "Guoqing Du", "Xiaoru Hu", "Hangyu Mao", "Ziyue Li", "Xingyu Zeng", "Rui Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0a2152fdc853957b6ebe9e2b4c70c1c5c579b675246fbfd82c3f4b3948bc4d1", "submitted_at": "2023-11-19T12:37:30Z", "updated_at": "2023-11-19T12:37:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.11315v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2311.11315", "identifiers": {"arxiv": "2311.11315"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.11315"}], "name": "TPTU-v2: Boosting Task Planning and Tool Usage of Large Language Model-based Agents in Real-world Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-19", "url": "https://arxiv.org/abs/2311.11315"}
{"aliases": [], "authors": ["Yu Tian", "Xiao Yang", "Jingyuan Zhang", "Yinpeng Dong", "Hang Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df779dcc767b085fec4500f7461c9909a941279c3a31fe3e2efd5f775170a61e", "submitted_at": "2023-11-20T15:50:09Z", "updated_at": "2024-02-02T08:28:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.11855v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2311.11855", "identifiers": {"arxiv": "2311.11855"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.11855"}], "name": "Evil Geniuses: Delving into the Safety of LLM-based Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-02", "url": "https://arxiv.org/abs/2311.11855"}
{"aliases": [], "authors": ["David Rein", "Betty Li Hou", "Asa Cooper Stickland", "Jackson Petty", "Richard Yuanzhe Pang", "Julien Dirani", "Julian Michael", "Samuel R. Bowman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d7927e2ce748f8cf939fe1c7f97dc0adf9a22e8c2f4e6fbe4895d31c78e4d09", "submitted_at": "2023-11-20T18:57:34Z", "updated_at": "2023-11-20T18:57:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.12022v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T08:01:28Z", "role": "primary", "sha256": "822cec31fa0db7ffa390fda806d88c2f9c0fe4d328742e7008b662916e92373a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2005.07648%2C2206.04615%2C2311.12022%2C2311.16922%2C2502.13130%2C2506.21506%2C2511.11332&max_results=7"}], "id": "arxiv:2311.12022", "identifiers": {"arxiv": "2311.12022"}, "kind": "paper", "last_verified_at": "2026-10-18T08:01:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.12022"}], "name": "GPQA: A Graduate-Level Google-Proof Q&A Benchmark", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-20", "url": "https://arxiv.org/abs/2311.12022"}
{"aliases": [], "authors": ["Mariia Mykhailova"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2e9f519c737eff3384333269d68c6b78c0236d634c87566cab93389dd45eb5d8", "submitted_at": "2023-11-21T19:55:23Z", "updated_at": "2023-12-15T23:37:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.12960v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2311.12960", "identifiers": {"arxiv": "2311.12960"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.12960"}], "name": "Teaching Quantum Computing using Microsoft Quantum Development Kit and Azure Quantum", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-15", "url": "https://arxiv.org/abs/2311.12960"}
{"aliases": [], "authors": ["Grégoire Mialon", "Clémentine Fourrier", "Craig Swift", "Thomas Wolf", "Yann LeCun", "Thomas Scialom"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29ddf2b7b4d495ee18bbd263cd77f86a71b3014b58a6f140c3bfce7cbae9c922", "submitted_at": "2023-11-21T20:34:47Z", "updated_at": "2023-11-21T20:34:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.12983v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2311.12983", "identifiers": {"arxiv": "2311.12983"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2311.12983"}], "name": "GAIA: a benchmark for General AI Assistants", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-11-21", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces questions that test assistant reasoning, multimodal understanding, browsing, and tool use beyond screen-action execution alone.", "topics": ["evaluation", "tool-integration"], "updated_at": "2023-11-21", "url": "https://arxiv.org/abs/2311.12983"}
{"aliases": [], "authors": ["Martin Zeppenfeld"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "272c4f18dacf1a16c6d0e21e122259d6d125affbf0fe7b178631813b5dc55dff", "submitted_at": "2023-11-23T18:04:56Z", "updated_at": "2023-11-23T18:04:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.14133v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2311.14133", "identifiers": {"arxiv": "2311.14133"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.14133"}], "name": "A robust framework for quantum computation using quasi-hidden molecular degrees of freedom", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-23", "url": "https://arxiv.org/abs/2311.14133"}
{"aliases": [], "authors": ["Jason Z. Kim", "Bart Larsen", "Linden Parkes"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1d07a8454a41c526056d99f5ad385cd5e8ee92384de27ed5f840f54253d03b0c", "submitted_at": "2023-11-27T06:45:31Z", "updated_at": "2023-11-27T06:45:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.15572v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2311.15572", "identifiers": {"arxiv": "2311.15572"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.15572"}], "name": "Shaping dynamical neural computations using spatiotemporal constraints", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-27", "url": "https://arxiv.org/abs/2311.15572"}
{"aliases": [], "authors": ["Sicong Leng", "Hang Zhang", "Guanzheng Chen", "Xin Li", "Shijian Lu", "Chunyan Miao", "Lidong Bing"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b3e3157e82a8ebe381a8fd2f5095ba37af60bfc5f5a700002bf75bbe4f4feba1", "submitted_at": "2023-11-28T16:26:35Z", "updated_at": "2023-11-28T16:26:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.16922v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T08:01:28Z", "role": "primary", "sha256": "822cec31fa0db7ffa390fda806d88c2f9c0fe4d328742e7008b662916e92373a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2005.07648%2C2206.04615%2C2311.12022%2C2311.16922%2C2502.13130%2C2506.21506%2C2511.11332&max_results=7"}], "id": "arxiv:2311.16922", "identifiers": {"arxiv": "2311.16922"}, "kind": "paper", "last_verified_at": "2026-10-18T08:01:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.16922"}], "name": "Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-28", "url": "https://arxiv.org/abs/2311.16922"}
{"aliases": [], "authors": ["Xin Liu", "Yichen Zhu", "Jindong Gu", "Yunshi Lan", "Chao Yang", "Yu Qiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b793b0158fc262b9e3cd2303f48b1381c1544c659352d124b0b6de6ffb72615c", "submitted_at": "2023-11-29T12:49:45Z", "updated_at": "2024-06-19T08:32:14Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2311.17600v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2311.17600", "identifiers": {"arxiv": "2311.17600"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.17600"}], "name": "MM-SafetyBench: A Benchmark for Safety Evaluation of Multimodal Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-19", "url": "https://arxiv.org/abs/2311.17600"}
{"aliases": [], "authors": ["Farbin Fayza", "Cansu Demirkiran", "Hanning Chen", "Che-Kai Liu", "Avi Mohan", "Hamza Errahmouni", "Sanggeon Yun", "Mohsen Imani", "David Zhang", "Darius Bunandar", "Ajay Joshi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c4d5c0a824e15460a6cb0513aef31be70e056105ffafce8285f3128b6a0421b8", "submitted_at": "2023-11-29T16:51:21Z", "updated_at": "2024-10-04T15:26:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.17801v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2311.17801", "identifiers": {"arxiv": "2311.17801"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.17801"}], "name": "Towards Efficient Hyperdimensional Computing Using Photonics", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-04", "url": "https://arxiv.org/abs/2311.17801"}
{"aliases": [], "authors": ["Constantin Chaumet", "Jakob Rehof"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "beb41574eb691fc4c02ba93e8494aa48f4aabd867d4bc2736635e220322761cf", "submitted_at": "2023-11-30T12:07:27Z", "updated_at": "2023-11-30T12:07:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.18492v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2311.18492", "identifiers": {"arxiv": "2311.18492"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.18492"}], "name": "CLS-CAD: Synthesizing CAD Assemblies in Fusion 360", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-11-30", "url": "https://arxiv.org/abs/2311.18492"}
{"aliases": [], "authors": ["Shin-Cheng Mu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37d2cb8816e3597e1aae87da684cf4568a5ff15d05e7a8ea02ccb86e1e6f7e27", "submitted_at": "2023-11-30T13:08:46Z", "updated_at": "2024-03-04T06:15:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.18528v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2311.18528", "identifiers": {"arxiv": "2311.18528"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.18528"}], "name": "Bottom-up computation using trees of sublists (Functional Pearl)", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-04", "url": "https://arxiv.org/abs/2311.18528"}
{"aliases": [], "authors": ["Hiroki Furuta", "Yutaka Matsuo", "Aleksandra Faust", "Izzeddin Gur"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4b00e0a9c74b7ac8819bdda540a1069c33cbac45f0c7fb853dee995f2feb9a59", "submitted_at": "2023-11-30T17:50:47Z", "updated_at": "2024-12-31T04:24:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2311.18751v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2311.18751", "identifiers": {"arxiv": "2311.18751"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2311.18751"}], "name": "Exposing Limitations of Language Model Agents in Sequential-Task Compositions on the Web", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-11-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "CompWoB adds 50 composed web-automation tasks to study generalization from simpler interactions. The paper compares prompted and fine-tuned agents, balances task training data, and probes sensitivity to changes in instruction order.", "topics": ["evaluation", "long-horizon", "robustness", "web-navigation"], "updated_at": "2024-12-31", "url": "https://arxiv.org/abs/2311.18751"}
{"aliases": [], "authors": ["Yongliang Shen", "Kaitao Song", "Xu Tan", "Wenqi Zhang", "Kan Ren", "Siyu Yuan", "Weiming Lu", "Dongsheng Li", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7751175e3fc55aa0422c73050eff40a7cc27b2a3ae9853be0359a6ae1e2a2af4", "submitted_at": "2023-11-30T18:02:44Z", "updated_at": "2024-11-01T14:37:37Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2311.18760v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2311.18760", "identifiers": {"arxiv": "2311.18760"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2311.18760"}], "name": "TaskBench: Benchmarking Large Language Models for Task Automation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-11-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-01", "url": "https://arxiv.org/abs/2311.18760"}
{"aliases": [], "authors": ["Penglin Cai", "Chi Zhang", "Yuhui Fu", "Haoqi Yuan", "Zongqing Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "58b723b5f33af6f2b82f58470a763db5e942fcd9b4324d9e1412b912ccc6449a", "submitted_at": "2023-12-05T06:00:52Z", "updated_at": "2025-12-26T06:02:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.02519v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2312.02519", "identifiers": {"arxiv": "2312.02519"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2312.02519"}], "name": "Creative Agents: Empowering Agents with Imagination for Creative Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-12-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-26", "url": "https://arxiv.org/abs/2312.02519"}
{"aliases": [], "authors": ["Kiana Ehsani", "Tanmay Gupta", "Rose Hendrix", "Jordi Salvador", "Luca Weihs", "Kuo-Hao Zeng", "Kunal Pratap Singh", "Yejin Kim", "Winson Han", "Alvaro Herrasti", "Ranjay Krishna", "Dustin Schwenk", "Eli VanderBilt", "Aniruddha Kembhavi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d38bb3f1a7ee583a61f98af43c9fe4ccd85b1da172fcd11df9d48f56f1245b55", "submitted_at": "2023-12-05T18:59:45Z", "updated_at": "2024-08-07T18:11:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.02976v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2312.02976", "identifiers": {"arxiv": "2312.02976"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2312.02976"}], "name": "SPOC: Imitating Shortest Paths in Simulation Enables Effective Navigation and Manipulation in the Real World", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-12-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-07", "url": "https://arxiv.org/abs/2312.02976"}
{"aliases": [], "authors": ["Sunjae Lee", "Junyoung Choi", "Jungjae Lee", "Munim Hasan Wasi", "Hojun Choi", "Steven Y. Ko", "Sangeun Oh", "Insik Shin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9b6f10b17513f565a13e035802d8bd681b5b0d468ba000df1427a8d59591e2c", "submitted_at": "2023-12-04T06:13:35Z", "updated_at": "2024-10-16T08:15:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.03003v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2312.03003", "identifiers": {"arxiv": "2312.03003"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.03003"}], "name": "Explore, Select, Derive, and Recall: Augmenting LLM with Human-like Memory for Mobile Task Automation", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2023-12-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MobileGPT, which organizes mobile-app experience into reusable subtasks using an explore, select, derive, and recall cycle. Its memory supports learning task procedures and adapting them to new contexts.", "topics": ["mobile-navigation", "memory", "planning"], "updated_at": "2024-10-16", "url": "https://arxiv.org/abs/2312.03003"}
{"aliases": [], "authors": ["Yingqiang Ge", "Yujie Ren", "Wenyue Hua", "Shuyuan Xu", "Juntao Tan", "Yongfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "742c1c5f0add06a42d3df69bbcefa42baf6383b3af1961bc2a05bf8473df2019", "submitted_at": "2023-12-06T18:50:26Z", "updated_at": "2023-12-09T18:10:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.03815v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2312.03815", "identifiers": {"arxiv": "2312.03815"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2312.03815"}], "name": "LLM as OS, Agents as Apps: Envisioning AIOS, Agents and the AIOS-Agent Ecosystem", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-12-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-09", "url": "https://arxiv.org/abs/2312.03815"}
{"aliases": [], "authors": ["Yanchu Guan", "Dong Wang", "Zhixuan Chu", "Shiyu Wang", "Feiyue Ni", "Ruihua Song", "Longfei Li", "Jinjie Gu", "Chenyi Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "598c8c30348d454c638944c932dd55355b5d3ecda6eb6f3910a27fffcab9622a", "submitted_at": "2023-12-04T07:51:58Z", "updated_at": "2023-12-04T07:51:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.06677v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2312.06677", "identifiers": {"arxiv": "2312.06677"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.06677"}], "name": "Intelligent Virtual Assistants with LLM-based Process Automation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The abstract names Alipay but does not establish Android or iOS specifically.", "published_at": "2023-12-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents LLMPA, a mobile process-automation architecture that decomposes requests, identifies interface elements, predicts actions and checks errors. Demonstrations evaluate multistep operations in Alipay.", "topics": ["mobile-navigation", "planning", "automation"], "updated_at": "2023-12-04", "url": "https://arxiv.org/abs/2312.06677"}
{"aliases": [], "authors": ["Zhongyi Han", "Guanglin Zhou", "Rundong He", "Jindong Wang", "Tailin Wu", "Yilong Yin", "Salman Khan", "Lina Yao", "Tongliang Liu", "Kun Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c126a68b4c5e2248289fce10b3c33bd2b0a56cc42c629546b51cd99b8f34e03", "submitted_at": "2023-12-12T16:48:07Z", "updated_at": "2024-02-25T08:10:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.07424v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2312.07424", "identifiers": {"arxiv": "2312.07424"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2312.07424"}], "name": "How Well Does GPT-4V(ision) Adapt to Distribution Shifts? A Preliminary Investigation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-12-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-25", "url": "https://arxiv.org/abs/2312.07424"}
{"aliases": [], "authors": ["Wenyi Hong", "Weihan Wang", "Qingsong Lv", "Jiazheng Xu", "Wenmeng Yu", "Junhui Ji", "Yan Wang", "Zihan Wang", "Yuxuan Zhang", "Juanzi Li", "Bin Xu", "Yuxiao Dong", "Ming Ding", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755198d83a3af9fe6d032f1446ec9bf57bf7e6f314d37787ce8a6aaa16bddc47", "submitted_at": "2023-12-14T13:20:57Z", "updated_at": "2024-12-27T06:56:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.08914v3"}, "publication": {"authors": ["Wenyi Hong", "Weihan Wang", "Qingsong Lv", "Jiazheng Xu", "Wenmeng Yu", "Junhui Ji", "Yan Wang", "Zihan Wang", "Yuxiao Dong", "Ming Ding", "Jie Tang"], "date_basis": "Year from publisher citation metadata and month from the same page's embedded BibTeX; no day is supplied.", "date_parts": [2024, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html", "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher page directly links this arXiv identifier. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. The publisher PDF, page 1, and landing page list eleven authors; the retained arXiv snapshot lists fourteen. Each edition retains its own authors.", "pages": "14281-14290", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf", "publisher": null, "series": null, "title": "CogAgent: A Visual Language Model for GUI Agents", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition", "volume": null}}, "environments": ["desktop", "android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:00Z", "role": "primary", "sha256": "43009fbc322dbe4f6cd1be01b7f414b4747ddcadaf4b63ae16f8eecc7ca6837b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:11.606232+00:00", "role": "primary", "sha256": "b9bfea178ee72b694eb575c73e3d7d8209a72706c9dda04296a70e17b0c98cf4", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2312.08914", "identifiers": {"arxiv": "2312.08914"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.08914"}, {"rel": "publication", "url": "https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html"}, {"rel": "publication_pdf", "url": "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf"}], "name": "CogAgent: A Visual Language Model for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-12-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a visual-language model with low- and high-resolution image encoders for screenshot understanding and GUI navigation.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-12-27", "url": "https://arxiv.org/abs/2312.08914"}
{"aliases": [], "authors": ["David Nordlund", "Zheng Chen", "Erik G. Larsson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61137e3878a75ddba6fe380f7cd4737588839b4236ac08c1c771e86446b8f472", "submitted_at": "2023-12-14T14:00:13Z", "updated_at": "2023-12-14T14:00:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.08950v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2312.08950", "identifiers": {"arxiv": "2312.08950"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2312.08950"}], "name": "Detecting Active Attacks in Over-the-Air Computation using Dummy Samples", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-12-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-14", "url": "https://arxiv.org/abs/2312.08950"}
{"aliases": [], "authors": ["Hao Li", "Xue Yang", "Zhaokai Wang", "Xizhou Zhu", "Jie Zhou", "Yu Qiao", "Xiaogang Wang", "Hongsheng Li", "Lewei Lu", "Jifeng Dai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff744dda42f804d5374f2f2205b34f80f38a419a79d5e4cae9443fca2f56e0e1", "submitted_at": "2023-12-14T18:58:12Z", "updated_at": "2024-03-30T15:35:16Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.09238v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}], "id": "arxiv:2312.09238", "identifiers": {"arxiv": "2312.09238"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2312.09238"}], "name": "Auto MC-Reward: Automated Dense Reward Design with Large Language Models for Minecraft", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-12-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-30", "url": "https://arxiv.org/abs/2312.09238"}
{"aliases": [], "authors": ["Mahmoud Jahanshahi", "Audris Mockus"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6908c92138adb51e53ded75650ef9c825d9ae71f79baca6b9121546b3fe01be7", "submitted_at": "2023-12-14T22:08:09Z", "updated_at": "2023-12-14T22:08:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.09370v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2312.09370", "identifiers": {"arxiv": "2312.09370"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2312.09370"}], "name": "Dataset: Copy-based Reuse in Open Source Software", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-12-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2023-12-14", "url": "https://arxiv.org/abs/2312.09370"}
{"aliases": [], "authors": ["Yunpeng Song", "Yiheng Bian", "Yongtao Tang", "Guiyu Ma", "Zhongmin Cai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df1390eab35db18dedaf135ddbbc9fdfcacbc9a70c51ec2609cf331cc05e06bd", "submitted_at": "2023-12-18T13:35:46Z", "updated_at": "2024-07-29T14:52:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.11190v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2312.11190", "identifiers": {"arxiv": "2312.11190"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.11190"}], "name": "VisionTasker: Mobile Task Automation Using Vision Based UI Understanding and LLM Task Planning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-12-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "VisionTasker converts screenshots into textual UI interpretations and presents one interface at a time to an LLM planner for action selection. It evaluates visual representations and execution of 147 tasks on an Android smartphone, including integration with programming by demonstration.", "topics": ["screen-understanding", "planning", "mobile-navigation"], "updated_at": "2024-07-29", "url": "https://arxiv.org/abs/2312.11190"}
{"aliases": [], "authors": ["Megan Kinniment", "Lucas Jun Koba Sato", "Haoxing Du", "Brian Goodrich", "Max Hasin", "Lawrence Chan", "Luke Harold Miles", "Tao R. Lin", "Hjalmar Wijk", "Joel Burget", "Aaron Ho", "Elizabeth Barnes", "Paul Christiano"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "998715dd4203d6c1254c773975320f7105614b0359a1eaf845626a0fbca15106", "submitted_at": "2023-12-18T19:27:09Z", "updated_at": "2024-01-04T18:46:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.11671v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2312.11671", "identifiers": {"arxiv": "2312.11671"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2312.11671"}], "name": "Evaluating Language-Model Agents on Realistic Autonomous Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-12-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-01-04", "url": "https://arxiv.org/abs/2312.11671"}
{"aliases": [], "authors": ["Weiyu Ma", "Qirui Mi", "Yongcheng Zeng", "Xue Yan", "Yuqiao Wu", "Runji Lin", "Haifeng Zhang", "Jun Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2936678bcb780881e7b65d7a08f9da479e9bf2841b83212dc769c6ed6d767a0f", "submitted_at": "2023-12-19T05:27:16Z", "updated_at": "2024-06-18T03:07:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.11865v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2312.11865", "identifiers": {"arxiv": "2312.11865"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2312.11865"}], "name": "Large Language Models Play StarCraft II: Benchmarks and A Chain of Summarization Approach", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2023-12-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-18", "url": "https://arxiv.org/abs/2312.11865"}
{"aliases": [], "authors": ["Difei Gao", "Lei Ji", "Zechen Bai", "Mingyu Ouyang", "Peiran Li", "Dongxing Mao", "Qinchen Wu", "Weichen Zhang", "Peiyi Wang", "Xiangwu Guo", "Hengxu Wang", "Luowei Zhou", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c771b076044c170728041a4f54662d21493adcc3742d40db08cb2f9bd0614655", "submitted_at": "2023-12-20T15:28:38Z", "updated_at": "2024-01-01T14:26:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.13108v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2312.13108", "identifiers": {"arxiv": "2312.13108"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.13108"}], "name": "ASSISTGUI: Task-Oriented Desktop Graphical User Interface Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-12-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Windows application tasks and an actor-critic agent framework for mouse-and-keyboard automation of desktop workflows.", "topics": ["evaluation", "planning"], "updated_at": "2024-01-01", "url": "https://arxiv.org/abs/2312.13108"}
{"aliases": [], "authors": ["Chi Zhang", "Zhao Yang", "Jiaxuan Liu", "Yanda Li", "Yucheng Han", "Xin Chen", "Zebiao Huang", "Bin Fu", "Gang Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61e54d83304755ce89c2ab69f7bb93a7fdc4d68cb5182daad9b5210f809d7ae1", "submitted_at": "2023-12-21T11:52:45Z", "updated_at": "2026-07-05T07:51:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.13771v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2312.13771", "identifiers": {"arxiv": "2312.13771"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.13771"}], "name": "AppAgent: Multimodal Agents as Smartphone Users", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-12-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds app knowledge through exploration or human demonstrations and uses it to guide multimodal smartphone interactions.", "topics": ["mobile-navigation", "memory"], "updated_at": "2026-07-05", "url": "https://arxiv.org/abs/2312.13771"}
{"aliases": [], "authors": ["Qi Chen", "Dileepa Pitawela", "Chongyang Zhao", "Gengze Zhou", "Hsiang-Ting Chen", "Qi Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47dedddced8f06bf436ab5a101d1ec4ecf4380c710e457dc7a35fc86d2e1f230", "submitted_at": "2023-12-25T22:13:26Z", "updated_at": "2023-12-25T22:13:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.15820v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2312.15820", "identifiers": {"arxiv": "2312.15820"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.15820"}], "name": "WebVLN: Vision-and-Language Navigation on Websites", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-12-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebVLN trains website navigation from question-based instructions using both rendered visual information and underlying HTML. It introduces WebVLN-v1 and a website-aware navigation network to study this digital variant of vision-and-language navigation.", "topics": ["web-navigation", "data-generation", "evaluation"], "updated_at": "2023-12-25", "url": "https://arxiv.org/abs/2312.15820"}
{"aliases": [], "authors": ["Lihang Pan", "Bowen Wang", "Chun Yu", "Yuxuan Chen", "Xiangyu Zhang", "Yuanchun Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f592712517da90bfeb807ab445034c3974b406e0ed6b4933fbd19078984b9b8", "submitted_at": "2023-12-26T14:20:36Z", "updated_at": "2023-12-26T14:20:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.16062v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2312.16062", "identifiers": {"arxiv": "2312.16062"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.16062"}], "name": "AutoTask: Executing Arbitrary Voice Commands by Exploring and Learning from Mobile GUI", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2023-12-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Executes voice-requested mobile tasks through GUI exploration, backtracking, and accumulated interaction experience.", "topics": ["mobile-navigation", "memory"], "updated_at": "2023-12-26", "url": "https://arxiv.org/abs/2312.16062"}
{"aliases": [], "authors": ["Boyuan Zheng", "Boyu Gou", "Jihyung Kil", "Huan Sun", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a005fb14b21cf17521f4ea966e555d9963092b709bd9280c6e72a9bf0f5fa93a", "submitted_at": "2024-01-03T08:33:09Z", "updated_at": "2024-03-12T23:14:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.01614v2"}, "publication": {"authors": ["Boyuan Zheng", "Boyu Gou", "Jihyung Kil", "Huan Sun", "Yu Su"], "date_basis": "Explicit publisher metadata: 2024/07/08; precision=day.", "date_parts": [2024, 7, 8], "doi": null, "event_dates": "21--27 Jul 2024", "evidence_urls": ["https://proceedings.mlr.press/v235/zheng24e.html"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "61349-61385", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/zheng24e/zheng24e.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "GPT-4V(ision) is a Generalist Web Agent, if Grounded", "type": "conference-paper", "url": "https://proceedings.mlr.press/v235/zheng24e.html", "venue": "Proceedings of the 41st International Conference on Machine Learning", "volume": "235"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:47+00:00", "role": "primary", "sha256": "eb144b5dfe51e5a9092d273fcffb85343808d011b03d22e059c9cc65e642ef33", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v235/zheng24e.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2401.01614", "identifiers": {"arxiv": "2401.01614"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.01614"}, {"rel": "publication", "url": "https://proceedings.mlr.press/v235/zheng24e.html"}, {"rel": "publication_pdf", "url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/zheng24e/zheng24e.pdf"}], "name": "GPT-4V(ision) is a Generalist Web Agent, if Grounded", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-01-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SeeAct and live-website evaluation, examining how visual planning and element grounding affect web-agent task completion.", "topics": ["web-navigation", "grounding"], "updated_at": "2024-03-12", "url": "https://arxiv.org/abs/2401.01614"}
{"aliases": [], "authors": ["Tinghe Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b8aaa2bccafd7f41802e5b7c25bc98f78307f77aeb36b7c75c57ef0f8e583a1", "submitted_at": "2024-01-04T03:44:42Z", "updated_at": "2024-01-17T06:35:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2401.04124v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2401.04124", "identifiers": {"arxiv": "2401.04124"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.04124"}], "name": "MobileAgent: enhancing mobile control via human-machine interaction and SOP integration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-01-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines operating-procedure context with user interaction to guide mobile task automation and handle personalized or sensitive requests.", "topics": ["mobile-navigation", "workflow", "human-in-the-loop"], "updated_at": "2024-01-17", "url": "https://arxiv.org/abs/2401.04124"}
{"aliases": [], "authors": ["Sivaprasad Omanakuttan", "Vikas Buchemmavari", "Jonathan A. Gross", "Ivan H Deutsch", "Milad Marvian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5ca2bf9307655e4cf7acac9e17ad3ba5eab1c065b10d12caa4f5c8bb7035f1dc", "submitted_at": "2024-01-08T22:56:05Z", "updated_at": "2024-06-11T18:18:00Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2401.04271v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2401.04271", "identifiers": {"arxiv": "2401.04271"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.04271"}], "name": "Fault-tolerant quantum computation using large spin cat-codes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-11", "url": "https://arxiv.org/abs/2401.04271"}
{"aliases": [], "authors": ["Yuanchun Li", "Hao Wen", "Weijun Wang", "Xiangyu Li", "Yizhen Yuan", "Guohong Liu", "Jiacheng Liu", "Wenxing Xu", "Xiang Wang", "Yi Sun", "Rui Kong", "Yile Wang", "Hanfei Geng", "Jian Luan", "Xuefeng Jin", "Zilong Ye", "Guanjing Xiong", "Fan Zhang", "Xiang Li", "Mengwei Xu", "Zhijun Li", "Peng Li", "Yang Liu", "Ya-Qin Zhang", "Yunxin Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc28f7eef225940b41b01940ae3ac594ca76ef4c6db5f19a99a4ac2519b95432", "submitted_at": "2024-01-10T09:25:45Z", "updated_at": "2024-05-08T06:16:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.05459v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2401.05459", "identifiers": {"arxiv": "2401.05459"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.05459"}], "name": "Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-08", "url": "https://arxiv.org/abs/2401.05459"}
{"aliases": [], "authors": ["Xueyu Hu", "Ziyu Zhao", "Shuang Wei", "Ziwei Chai", "Qianli Ma", "Guoyin Wang", "Xuwu Wang", "Jing Su", "Jingjing Xu", "Ming Zhu", "Yao Cheng", "Jianbo Yuan", "Jiwei Li", "Kun Kuang", "Yang Yang", "Hongxia Yang", "Fei Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b9e442e51a9d16274f0e323f255bb25a34384c1770d5103bf9c18cea08c4c30", "submitted_at": "2024-01-10T19:04:00Z", "updated_at": "2024-03-11T07:57:59Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2401.05507v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2401.05507", "identifiers": {"arxiv": "2401.05507"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.05507"}], "name": "InfiAgent-DABench: Evaluating Agents on Data Analysis Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-11", "url": "https://arxiv.org/abs/2401.05507"}
{"aliases": [], "authors": ["Shuai Zhao", "Meihuizi Jia", "Luu Anh Tuan", "Fengjun Pan", "Jinming Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57bc41f98762fb81deea568b113d21f86dbdc7ceef7d979841fd2f9a3b8b8dd4", "submitted_at": "2024-01-11T14:38:19Z", "updated_at": "2024-10-09T11:46:24Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2401.05949v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2401.05949", "identifiers": {"arxiv": "2401.05949"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.05949"}], "name": "Universal Vulnerabilities in Large Language Models: Backdoor Attacks for In-context Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-09", "url": "https://arxiv.org/abs/2401.05949"}
{"aliases": [], "authors": ["Yi Zeng", "Hongpeng Lin", "Jingwen Zhang", "Diyi Yang", "Ruoxi Jia", "Weiyan Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "feacd6d6fd99c426b70a6540db82ccc7d03c56e9ce22346cf3684375c20b8e36", "submitted_at": "2024-01-12T16:13:24Z", "updated_at": "2024-01-23T22:46:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.06373v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2401.06373", "identifiers": {"arxiv": "2401.06373"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.06373"}], "name": "How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-01-23", "url": "https://arxiv.org/abs/2401.06373"}
{"aliases": [], "authors": ["Jay Zhangjie Wu", "Guian Fang", "Haoning Wu", "Xintao Wang", "Yixiao Ge", "Xiaodong Cun", "David Junhao Zhang", "Jia-Wei Liu", "Yuchao Gu", "Rui Zhao", "Weisi Lin", "Wynne Hsu", "Ying Shan", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33e68f587d25577d4afd2ebd33fb04bc2ed1d5e1435075376b78896a78855422", "submitted_at": "2024-01-15T15:42:39Z", "updated_at": "2024-01-15T15:42:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2401.07781v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2401.07781", "identifiers": {"arxiv": "2401.07781"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.07781"}], "name": "Towards A Better Metric for Text-to-Video Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-01-15", "url": "https://arxiv.org/abs/2401.07781"}
{"aliases": [], "authors": ["Tongxin Yuan", "Zhiwei He", "Lingzhong Dong", "Yiming Wang", "Ruijie Zhao", "Tian Xia", "Lizhen Xu", "Binglin Zhou", "Fangqi Li", "Zhuosheng Zhang", "Rui Wang", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d9aeab527f85fd64aab821f5440967fad5350b70273a09c79463fe7959671228", "submitted_at": "2024-01-18T14:40:46Z", "updated_at": "2024-10-05T06:50:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2401.10019v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2401.10019", "identifiers": {"arxiv": "2401.10019"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.10019"}], "name": "R-Judge: Benchmarking Safety Risk Awareness for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-05", "url": "https://arxiv.org/abs/2401.10019"}
{"aliases": [], "authors": ["Kanzhi Cheng", "Qiushi Sun", "Yougang Chu", "Fangzhi Xu", "Yantao Li", "Jianbing Zhang", "Zhiyong Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c92ec9323eec9362aa07984f9fb1cbfbe98ecae426a9f8adc9d436328a87aed3", "submitted_at": "2024-01-17T08:10:35Z", "updated_at": "2024-02-23T04:36:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.10935v2"}, "publication": {"authors": ["Kanzhi Cheng", "Qiushi Sun", "Yougang Chu", "Fangzhi Xu", "Yantao Li", "Jianbing Zhang", "Zhiyong Wu"], "date_basis": "Explicit publisher metadata: 2024/8; precision=month.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.505", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.505/", "https://aclanthology.org/2024.acl-long.505.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The ACL landing metadata spells the fifth author Li YanTao. The formal PDF title page spells Yantao Li; the publication author field uses the PDF. The retained arXiv author list is untouched. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "9313-9332", "pdf_url": "https://aclanthology.org/2024.acl-long.505.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.505/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "62d56e417487461c8b34da15167eafd5f14f58092e4269222dc5c1c9048731a9", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.acl-long.505/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:56:04+00:00", "role": "primary", "sha256": "8e1f6809e35751d762714286c55e79e81bae071608f90ef561d804aa59d7e406", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2024.acl-long.505.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2401.10935", "identifiers": {"arxiv": "2401.10935"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.10935"}, {"rel": "publication", "url": "https://aclanthology.org/2024.acl-long.505/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.acl-long.505.pdf"}], "name": "SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-01-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains a screenshot-driven GUI agent with grounding pretraining and introduces ScreenSpot for evaluating interface target localization.", "topics": ["grounding", "data-generation"], "updated_at": "2024-02-23", "url": "https://arxiv.org/abs/2401.10935"}
{"aliases": [], "authors": ["Zaibin Zhang", "Yongting Zhang", "Lijun Li", "Hongzhi Gao", "Lijun Wang", "Huchuan Lu", "Feng Zhao", "Yu Qiao", "Jing Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67df2290cac3567713de0eca43d58eade547f312ae82a3332eb545050c262c24", "submitted_at": "2024-01-22T12:11:55Z", "updated_at": "2024-08-20T06:45:50Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2401.11880v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2401.11880", "identifiers": {"arxiv": "2401.11880"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.11880"}], "name": "PsySafe: A Comprehensive Framework for Psychological-based Attack, Defense, and Evaluation of Multi-agent System Safety", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-20", "url": "https://arxiv.org/abs/2401.11880"}
{"aliases": [], "authors": ["Zhen Xiang", "Fengqing Jiang", "Zidi Xiong", "Bhaskar Ramasubramanian", "Radha Poovendran", "Bo Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2294dcef3c75e7af37fbac6edfe91cf95bd2a61dd3bdb3f04e183406d6294bf5", "submitted_at": "2024-01-20T04:53:35Z", "updated_at": "2024-01-20T04:53:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2401.12242v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T11:04:43Z", "role": "primary", "sha256": "0901e063a911164dd291042946a9a0bc63b53d61da59666259c34fb2560e17e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1610.08401%2C2005.03823%2C2005.09161%2C2007.08745%2C2112.09332%2C2205.00445%2C2205.12255%2C2205.12700%2C2206.05751%2C2302.03668%2C2304.07590%2C2304.08354%2C2305.00944%2C2305.02424%2C2305.10847%2C2305.13300%2C2305.14710%2C2305.14950%2C2305.15334%2C2306.06624%2C2306.13213%2C2306.15447%2C2306.17194%2C2306.17882%2C2307.07924%2C2307.10490%2C2307.14539%2C2307.15043%2C2308.00352%2C2308.03427%2C2308.08155%2C2309.06055%2C2309.10253%2C2309.15817%2C2310.02446%2C2310.06387%2C2310.10077%2C2311.09447%2C2311.11315%2C2401.05949%2C2401.06373%2C2401.10019%2C2401.12242&max_results=43"}], "id": "arxiv:2401.12242", "identifiers": {"arxiv": "2401.12242"}, "kind": "paper", "last_verified_at": "2026-10-18T11:04:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.12242"}], "name": "BadChain: Backdoor Chain-of-Thought Prompting for Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-01-20", "url": "https://arxiv.org/abs/2401.12242"}
{"aliases": [], "authors": ["Chang Ma", "Junlei Zhang", "Zhihao Zhu", "Cheng Yang", "Yujiu Yang", "Yaohui Jin", "Zhenzhong Lan", "Lingpeng Kong", "Junxian He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "22c5a60f9ed7a8488ec31fbf24d392e9674b0cae79ab1b269fd9d57202025801", "submitted_at": "2024-01-24T01:51:00Z", "updated_at": "2024-12-23T20:12:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.13178v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "topics", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:30Z", "role": "primary", "sha256": "bbfef23bd53d2febb258a91d5a706f630cd597909bdbea56f16ac97dbe4777c0", "supports": ["summary", "environments", "scope", "notes"], "url": "https://arxiv.org/html/2401.13178v2"}], "id": "arxiv:2401.13178", "identifiers": {"arxiv": "2401.13178"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.13178"}], "name": "AgentBoard: An Analytical Evaluation Board of Multi-turn LLM Agents", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The full paper documents search/click WebShop actions and 11 browser actions in WebArena, while the overall benchmark also covers non-web domains.", "published_at": "2024-01-24", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces AgentBoard, a multiturn agent evaluation framework with intermediate progress metrics and analysis tools. Its broader task suite includes WebShop and WebArena interactions with stepwise progress measurements.", "topics": ["evaluation", "observability", "web-navigation"], "updated_at": "2024-12-23", "url": "https://arxiv.org/abs/2401.13178"}
{"aliases": [], "authors": ["Jing Yu Koh", "Robert Lo", "Lawrence Jang", "Vikram Duvvur", "Ming Chong Lim", "Po-Yu Huang", "Graham Neubig", "Shuyan Zhou", "Ruslan Salakhutdinov", "Daniel Fried"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a705b87b43464120ebb01b28b0f6a9e31fad68f3d463c88354f2005dab940c3", "submitted_at": "2024-01-24T18:35:21Z", "updated_at": "2024-06-06T02:01:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.13649v2"}, "publication": {"authors": ["Jing Yu Koh", "Robert Lo", "Lawrence Jang", "Vikram Duvvur", "Ming Chong Lim", "Po-Yu Huang", "Graham Neubig", "Shuyan Zhou", "Ruslan Salakhutdinov", "Daniel Fried"], "date_basis": "Explicit publisher metadata: 2024/8; precision=month.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.50", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.50/", "https://aclanthology.org/2024.acl-long.50.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The formal PDF title adds Visually Grounded; the landing metadata uses Visual Web Tasks. PDF authors include Ming Chong Lim and Ruslan Salakhutdinov; the landing abbreviates them to Ming Lim and Russ Salakhutdinov. Publication title/authors use the PDF; the retained arXiv title remains unchanged. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "881-905", "pdf_url": "https://aclanthology.org/2024.acl-long.50.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "VisualWebArena: Evaluating Multimodal Agents on Realistic Visually Grounded Web Tasks", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.50/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "fc908d474b12a6c0f3ff3e48129d29c809e9f3a0365a3cf647b70273f266849a", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.acl-long.50/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:56:04+00:00", "role": "primary", "sha256": "6fc85c38c975c2af170c1d1a28f72ba557906d9019d1d9106a15215868bb491e", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://aclanthology.org/2024.acl-long.50.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2401.13649", "identifiers": {"arxiv": "2401.13649"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.13649"}, {"rel": "publication", "url": "https://aclanthology.org/2024.acl-long.50/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.acl-long.50.pdf"}], "name": "VisualWebArena: Evaluating Multimodal Agents on Realistic Visual Web Tasks", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-01-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces visually grounded website tasks for evaluating multimodal agents that must interpret both images and text.", "topics": ["evaluation", "grounding", "web-navigation"], "updated_at": "2024-06-06", "url": "https://arxiv.org/abs/2401.13649"}
{"aliases": [], "authors": ["Hongliang He", "Wenlin Yao", "Kaixin Ma", "Wenhao Yu", "Yong Dai", "Hongming Zhang", "Zhenzhong Lan", "Dong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dbc7121c9606c5987df0941b849c216ad8731bebcc74a061cf3f154379d7b8da", "submitted_at": "2024-01-25T03:33:18Z", "updated_at": "2024-06-06T18:37:34Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2401.13919v4"}, "publication": {"authors": ["Hongliang He", "Wenlin Yao", "Kaixin Ma", "Wenhao Yu", "Yong Dai", "Hongming Zhang", "Zhenzhong Lan", "Dong Yu"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.371", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.371/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "6864-6890", "pdf_url": "https://aclanthology.org/2024.acl-long.371.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.371/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:48Z", "role": "primary", "sha256": "9772713e428cc47e54a6a236aef9024e44c97c6794a746ade767b51bdba7679a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.acl-long.371/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2401.13919", "identifiers": {"arxiv": "2401.13919"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.13919"}, {"rel": "publication", "url": "https://aclanthology.org/2024.acl-long.371/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.acl-long.371.pdf"}], "name": "WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-01-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines multimodal web interaction with a real-website task collection and an automated evaluation protocol.", "topics": ["web-navigation", "evaluation"], "updated_at": "2024-06-06", "url": "https://arxiv.org/abs/2401.13919"}
{"aliases": [], "authors": ["Minh Duc Vu", "Han Wang", "Zhuang Li", "Jieshan Chen", "Shengdong Zhao", "Zhenchang Xing", "Chunyang Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a1c95264a931b10a57ee4b364498e2d2795abaf9a3e89096a0f29d2408174693", "submitted_at": "2024-01-25T16:02:56Z", "updated_at": "2024-08-14T00:48:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2401.14268v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2401.14268", "identifiers": {"arxiv": "2401.14268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.14268"}], "name": "GPTVoiceTasker: Advancing Multi-step Mobile Task Efficiency Through Dynamic Interface Exploration and Learning", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-01-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GPTVoiceTasker, a mobile assistant that interprets user commands, executes device interactions, and learns from prior commands to streamline later task execution. The work includes a user study of task efficiency.", "topics": ["mobile-navigation", "human-in-the-loop", "memory"], "updated_at": "2024-08-14", "url": "https://arxiv.org/abs/2401.14268"}
{"aliases": [], "authors": ["Enrico Picco", "Lina Jaurigue", "Kathy Lüdge", "Serge Massar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3288597edca7b77ba05d5c11a24732582385874404ee728bd5ebba40b25d8be0", "submitted_at": "2024-01-25T18:20:37Z", "updated_at": "2024-01-25T18:20:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2401.14371v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2401.14371", "identifiers": {"arxiv": "2401.14371"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.14371"}], "name": "Efficient Optimisation of Physical Reservoir Computers using only a Delayed Input", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-01-25", "url": "https://arxiv.org/abs/2401.14371"}
{"aliases": [], "authors": ["Junyang Wang", "Haiyang Xu", "Jiabo Ye", "Ming Yan", "Weizhou Shen", "Ji Zhang", "Fei Huang", "Jitao Sang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81ac041a3e5f98bfc23c91f768b32ad7adc3f28d8c9d55a285bb29906c8b5a9c", "submitted_at": "2024-01-29T13:46:37Z", "updated_at": "2024-04-18T06:53:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.16158v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2401.16158", "identifiers": {"arxiv": "2401.16158"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.16158"}], "name": "Mobile-Agent: Autonomous Multi-Modal Mobile Device Agent with Visual Perception", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-01-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses visual perception tools and task decomposition to operate mobile applications without relying on app interface metadata.", "topics": ["mobile-navigation", "planning"], "updated_at": "2024-04-18", "url": "https://arxiv.org/abs/2401.16158"}
{"aliases": [], "authors": ["Yelaman Abdullin", "Diego Molla-Aliod", "Bahadorreza Ofoghi", "John Yearwood", "Qingyang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d26818ee4126590263822c9d3a4195c8a525aaa8f462ec09db304f880ce65aba", "submitted_at": "2024-01-30T21:49:30Z", "updated_at": "2024-01-30T21:49:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2401.17461v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2401.17461", "identifiers": {"arxiv": "2401.17461"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2401.17461"}], "name": "Synthetic Dialogue Dataset Generation using LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-01-30", "url": "https://arxiv.org/abs/2401.17461"}
{"aliases": [], "authors": ["Sippo Rossi", "Alisia Marianne Michel", "Raghava Rao Mukkamala", "Jason Bennett Thatcher"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bad79072484c77286cb38923d10e13fdbe8abce2eeef6d5dfbd33d310e288170", "submitted_at": "2024-01-31T19:52:00Z", "updated_at": "2024-01-31T19:52:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.00898v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.00898", "identifiers": {"arxiv": "2402.00898"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.00898"}], "name": "An Early Categorization of Prompt Injection Attacks on Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-01-31", "url": "https://arxiv.org/abs/2402.00898"}
{"aliases": [], "authors": ["Xingyao Wang", "Yangyi Chen", "Lifan Yuan", "Yizhe Zhang", "Yunzhu Li", "Hao Peng", "Heng Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e053453816217e97c5909f22f49b67c2a007b69d1148761e0a83017200f035e", "submitted_at": "2024-02-01T21:38:58Z", "updated_at": "2024-06-07T01:53:07Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2402.01030v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.01030", "identifiers": {"arxiv": "2402.01030"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.01030"}], "name": "Executable Code Actions Elicit Better LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-07", "url": "https://arxiv.org/abs/2402.01030"}
{"aliases": [], "authors": ["Sihao Hu", "Tiansheng Huang", "Ling Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6c083a7e16da26e92352c8b3dd52eada0dec1ce074d491e8f8fc769aecbaa6c5", "submitted_at": "2024-02-02T03:22:12Z", "updated_at": "2024-04-02T15:46:35Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.01118v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2402.01118", "identifiers": {"arxiv": "2402.01118"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.01118"}], "name": "PokeLLMon: A Human-Parity Agent for Pokemon Battles with Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-02", "url": "https://arxiv.org/abs/2402.01118"}
{"aliases": [], "authors": ["Wenyue Hua", "Xianjun Yang", "Mingyu Jin", "Zelong Li", "Wei Cheng", "Ruixiang Tang", "Yongfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "39bb08001266f4c7f96d660a06b6fae3864a7774008db2772dad8efa1985ab9c", "submitted_at": "2024-02-02T17:26:23Z", "updated_at": "2024-10-03T22:12:05Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2402.01586v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.01586", "identifiers": {"arxiv": "2402.01586"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.01586"}], "name": "TrustAgent: Towards Safe and Trustworthy LLM-based Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-03", "url": "https://arxiv.org/abs/2402.01586"}
{"aliases": [], "authors": ["Yi Zhao", "Yilin Zhang", "Rong Xiang", "Jing Li", "Hillming Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f08bce4a8aca2e33d362b79da821106ebdad1f816f712a8a633a3a2be756b6e", "submitted_at": "2024-01-29T08:28:32Z", "updated_at": "2024-02-10T14:08:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.01735v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.01735", "identifiers": {"arxiv": "2402.01735"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.01735"}], "name": "VIALM: A Survey and Benchmark of Visually Impaired Assistance with Large Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-10", "url": "https://arxiv.org/abs/2402.01735"}
{"aliases": [], "authors": ["A. O. Dadukin", "N. I. Pchelintseva"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e81184cf28ec50b43c32a84ed61669313d03439b804f99f551a281fb67083d8", "submitted_at": "2024-01-07T15:37:56Z", "updated_at": "2024-01-07T15:37:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.03323v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.03323", "identifiers": {"arxiv": "2402.03323"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.03323"}], "name": "Portable medical devices creation technology by using the Bluetooth module", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-01-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-01-07", "url": "https://arxiv.org/abs/2402.03323"}
{"aliases": [], "authors": ["Daechul Ahn", "Yura Choi", "Youngjae Yu", "Dongyeop Kang", "Jonghyun Choi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e58de9044baa1c996b2297d0fb9a9cedc2e2f62621ad0adc7e155f81d92ec277", "submitted_at": "2024-02-06T06:27:40Z", "updated_at": "2024-06-17T08:20:33Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.03746v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2402.03746", "identifiers": {"arxiv": "2402.03746"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.03746"}], "name": "Tuning Large Multimodal Models for Videos using Reinforcement Learning from AI Feedback", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-17", "url": "https://arxiv.org/abs/2402.03746"}
{"aliases": [], "authors": ["Xiangru Tang", "Qiao Jin", "Kunlun Zhu", "Tongxin Yuan", "Yichi Zhang", "Wangchunshu Zhou", "Meng Qu", "Yilun Zhao", "Jian Tang", "Zhuosheng Zhang", "Arman Cohan", "Zhiyong Lu", "Mark Gerstein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1e55a6727104db03611d755b99c128f76ead329a2e40fa1759949c510feafa5e", "submitted_at": "2024-02-06T18:54:07Z", "updated_at": "2025-07-21T18:59:21Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2402.04247v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.04247", "identifiers": {"arxiv": "2402.04247"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.04247"}], "name": "Risks of AI Scientists: Prioritizing Safeguarding Over Autonomy", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-21", "url": "https://arxiv.org/abs/2402.04247"}
{"aliases": [], "authors": ["Jihyung Kil", "Chan Hee Song", "Boyuan Zheng", "Xiang Deng", "Yu Su", "Wei-Lun Chao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3da421fc440f7dcfc53d8ef218f55389e4da4dbac196f54b40402bc877ff0365", "submitted_at": "2024-02-06T23:52:10Z", "updated_at": "2024-03-30T05:18:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.04476v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.04476", "identifiers": {"arxiv": "2402.04476"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.04476"}], "name": "Dual-View Visual Contextualization for Web Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses screenshot geometry and visual content to enrich HTML element representations for web navigation. Nearby elements contribute textual and visual context, with evaluation on cross-task, cross-website, and cross-domain Mind2Web splits.", "topics": ["grounding", "screen-understanding", "web-navigation"], "updated_at": "2024-03-30", "url": "https://arxiv.org/abs/2402.04476"}
{"aliases": [], "authors": ["Gilles Baechler", "Srinivas Sunkara", "Maria Wang", "Fedir Zubach", "Hassan Mansoor", "Vincent Etter", "Victor Cărbune", "Jason Lin", "Jindong Chen", "Abhanshu Sharma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b293ab33fe0f3607eccc2b99cbf06fb459792444433a4179bb65e081f33e426c", "submitted_at": "2024-02-07T06:42:33Z", "updated_at": "2024-07-04T07:08:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.04615v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.04615", "identifiers": {"arxiv": "2402.04615"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.04615"}], "name": "ScreenAI: A Vision-Language Model for UI and Infographics Understanding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ScreenAI and screen-annotation training that supports UI question answering, navigation, and summarization data generation.", "topics": ["screen-understanding", "data-generation"], "updated_at": "2024-07-04", "url": "https://arxiv.org/abs/2402.04615"}
{"aliases": [], "authors": ["Yue Jiang", "Luis A. Leiva", "Paul R. B. Houssel", "Hamed R. Tavakoli", "Julia Kylmälä", "Antti Oulasvirta"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a1b1dc5d3541b11902569ff4a0d2248a2e1ca8e969d7630054ecd15a921c3f3", "submitted_at": "2024-02-07T19:15:30Z", "updated_at": "2024-02-07T19:15:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.05202v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.05202", "identifiers": {"arxiv": "2402.05202"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.05202"}], "name": "UEyes: An Eye-Tracking Dataset across User Interface Types", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-07", "url": "https://arxiv.org/abs/2402.05202"}
{"aliases": [], "authors": ["Xing Han Lù", "Zdeněk Kasner", "Siva Reddy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a7c1a1f260f9287e8999525e1adc5534a70f7b96271d469973fceee1ac23609", "submitted_at": "2024-02-08T18:58:02Z", "updated_at": "2024-09-10T22:22:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.05930v2"}, "publication": {"authors": ["Xing Han Lù", "Zdeněk Kasner", "Siva Reddy"], "date_basis": "Explicit publisher metadata: 2024/07/08; precision=day.", "date_parts": [2024, 7, 8], "doi": null, "event_dates": "21--27 Jul 2024", "evidence_urls": ["https://proceedings.mlr.press/v235/lu24e.html", "https://raw.githubusercontent.com/mlresearch/v235/main/assets/lu24e/lu24e.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The PMLR landing renders Xing Han Lu; the paper title page has Xing Han Lù. Publication authors use the PDF. The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "33007-33056", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/lu24e/lu24e.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "WebLINX: Real-World Website Navigation with Multi-Turn Dialogue", "type": "conference-paper", "url": "https://proceedings.mlr.press/v235/lu24e.html", "venue": "Proceedings of the 41st International Conference on Machine Learning", "volume": "235"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:47+00:00", "role": "primary", "sha256": "1b30008f833b3b06238d87d0509a10876a08153a78a06c79bc7590c30c471777", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v235/lu24e.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:46+00:00", "role": "primary", "sha256": "c644b1e3f071879aec8ea0e741e31e84193905dd730382ada54f1f24d04298ab", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/lu24e/lu24e.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.05930", "identifiers": {"arxiv": "2402.05930"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.05930"}, {"rel": "publication", "url": "https://proceedings.mlr.press/v235/lu24e.html"}, {"rel": "publication_pdf", "url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/lu24e/lu24e.pdf"}], "name": "WebLINX: Real-World Website Navigation with Multi-Turn Dialogue", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-02-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Collects expert demonstrations of conversational web navigation and studies retrieval-based element selection for processing large pages.", "topics": ["data-generation", "web-navigation", "evaluation"], "updated_at": "2024-09-10", "url": "https://arxiv.org/abs/2402.05930"}
{"aliases": [], "authors": ["Andrey Moskalenko", "Vlad Shakhuro", "Anna Vorontsova", "Anton Konushin", "Anton Antonov", "Alexander Krapukhin", "Denis Shepelev", "Konstantin Soshin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "58bd995e26bb75944daa3df40d630e215be940278b76d5277f265247714383c6", "submitted_at": "2024-02-09T01:36:21Z", "updated_at": "2024-02-09T01:36:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.06132v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.06132", "identifiers": {"arxiv": "2402.06132"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.06132"}], "name": "TETRIS: Towards Exploring the Robustness of Interactive Segmentation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-09", "url": "https://arxiv.org/abs/2402.06132"}
{"aliases": [], "authors": ["Mingzhe Xing", "Rongkai Zhang", "Hui Xue", "Qi Chen", "Fan Yang", "Zhen Xiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0504cfffc5d46ecf447a4fdeac561aede9482bb8640fcfd9d7f02cf6eaf47597", "submitted_at": "2024-02-09T18:19:25Z", "updated_at": "2024-02-09T18:19:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.06596v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.06596", "identifiers": {"arxiv": "2402.06596"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.06596"}], "name": "Understanding the Weakness of Large Language Model Agents within a Complex Android Environment", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AndroidArena and analyzes agent limitations in cross-application tasks, user constraints, exploration, and reflection.", "topics": ["evaluation", "cross-app"], "updated_at": "2024-02-09", "url": "https://arxiv.org/abs/2402.06596"}
{"aliases": [], "authors": ["Simone Parisi", "Montaser Mohammedalamen", "Alireza Kazemipour", "Matthew E. Taylor", "Michael Bowling"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "847686941923beb89ac0b5db351bbd0eb50067e874110fe426483f010a16e621", "submitted_at": "2024-02-09T23:00:29Z", "updated_at": "2024-02-13T19:49:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.06819v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.06819", "identifiers": {"arxiv": "2402.06819"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.06819"}], "name": "Monitored Markov Decision Processes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-13", "url": "https://arxiv.org/abs/2402.06819"}
{"aliases": [], "authors": ["Zhiyong Wu", "Chengcheng Han", "Zichen Ding", "Zhenmin Weng", "Zhoumianze Liu", "Shunyu Yao", "Tao Yu", "Lingpeng Kong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "674b005f670a3593bd215a8ff75eac05c90ee068578fc26afdd998c40f5b7b71", "submitted_at": "2024-02-12T07:29:22Z", "updated_at": "2024-02-15T09:30:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.07456v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.07456", "identifiers": {"arxiv": "2402.07456"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.07456"}], "name": "OS-Copilot: Towards Generalist Computer Agents with Self-Improvement", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Provides an operating-system agent framework connecting applications, files, terminals, and web tools, with accumulated skills for subsequent tasks.", "topics": ["tool-integration", "planning"], "updated_at": "2024-02-15", "url": "https://arxiv.org/abs/2402.07456"}
{"aliases": [], "authors": ["Chaoyun Zhang", "Liqun Li", "Shilin He", "Xu Zhang", "Bo Qiao", "Si Qin", "Minghua Ma", "Yu Kang", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "533acb1c9aa3f1c77adb2c67779da86e3d5de95264ba3ec1f81d803f67da7753", "submitted_at": "2024-02-08T15:40:35Z", "updated_at": "2024-05-23T05:18:58Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2402.07939v5"}, "publication": {"authors": ["Chaoyun Zhang", "Liqun Li", "Shilin He", "Xu Zhang", "Bo Qiao", "Si Qin", "Minghua Ma", "Yu Kang", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-long.26", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-long.26/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "597-622", "pdf_url": "https://aclanthology.org/2025.naacl-long.26.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "UFO: A UI-Focused Agent for Windows OS Interaction", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-long.26/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", "volume": null}}, "environments": ["windows", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:48Z", "role": "primary", "sha256": "bbe15e312bcbc923811ef73e561fa7db6a7244ee3747bb9e3792e166b9327235", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-long.26/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.07939", "identifiers": {"arxiv": "2402.07939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.07939"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-long.26/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-long.26.pdf"}], "name": "UFO: A UI-Focused Agent for Windows OS Interaction", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "UFO uses two agents to inspect Windows application interfaces and their control metadata, then grounds and executes user requests within and across applications. The study evaluates this UI-focused architecture on nine Windows applications.", "topics": ["cross-app", "planning", "grounding", "automation"], "updated_at": "2024-05-23", "url": "https://arxiv.org/abs/2402.07939"}
{"aliases": [], "authors": ["Runliang Niu", "Jindong Li", "Shiqi Wang", "Yali Fu", "Xiyu Hu", "Xueyuan Leng", "He Kong", "Yi Chang", "Qi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63424a86302625b15a91ff0252ca16c0a6448160bfd748ba32e386c82883584e", "submitted_at": "2024-02-09T02:33:45Z", "updated_at": "2024-02-09T02:33:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.07945v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.07945", "identifiers": {"arxiv": "2402.07945"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.07945"}], "name": "ScreenAgent: A Vision Language Model-driven Computer Control Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines a screenshot-and-action dataset with a planning, acting, and reflecting loop for computer control through mouse and keyboard actions.", "topics": ["automation", "planning"], "updated_at": "2024-02-09", "url": "https://arxiv.org/abs/2402.07945"}
{"aliases": [], "authors": ["Xiangming Gu", "Xiaosen Zheng", "Tianyu Pang", "Chao Du", "Qian Liu", "Ye Wang", "Jing Jiang", "Min Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c47a3db8231fd799c5c4586c03442dc63dbb78541fb528c522aed38b82059d7", "submitted_at": "2024-02-13T16:06:17Z", "updated_at": "2024-06-03T14:15:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.08567v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.08567", "identifiers": {"arxiv": "2402.08567"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.08567"}], "name": "Agent Smith: A Single Image Can Jailbreak One Million Multimodal LLM Agents Exponentially Fast", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-03", "url": "https://arxiv.org/abs/2402.08567"}
{"aliases": [], "authors": ["Yutaro Yamada", "Khyathi Chandu", "Yuchen Lin", "Jack Hessel", "Ilker Yildirim", "Yejin Choi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccdaf1ad4432ad417773eb0bbd78296d8ab48221b64f6b52328ce705f2c05007", "submitted_at": "2024-02-14T09:51:05Z", "updated_at": "2024-02-14T09:51:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.09052v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.09052", "identifiers": {"arxiv": "2402.09052"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.09052"}], "name": "L3GO: Language Agents with Chain-of-3D-Thoughts for Generating Unconventional Objects", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-14", "url": "https://arxiv.org/abs/2402.09052"}
{"aliases": [], "authors": ["Cheng Qian", "Bingxiang He", "Zhong Zhuang", "Jia Deng", "Yujia Qin", "Xin Cong", "Zhong Zhang", "Jie Zhou", "Yankai Lin", "Zhiyuan Liu", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45719b1e0fc841852300936065197c8af86a4bdef6d6dc5bcb59de603ce240fd", "submitted_at": "2024-02-14T14:36:30Z", "updated_at": "2024-02-15T09:59:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.09205v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2402.09205", "identifiers": {"arxiv": "2402.09205"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.09205"}], "name": "Tell Me More! Towards Implicit User Intention Understanding of Language Model Driven Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-15", "url": "https://arxiv.org/abs/2402.09205"}
{"aliases": [], "authors": ["Lingbo Mo", "Zeyi Liao", "Boyuan Zheng", "Yu Su", "Chaowei Xiao", "Huan Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6d81a6b9de6fa32c8f5749d029d0819ddf3b310ae00a4872877ee72580e8879", "submitted_at": "2024-02-15T18:51:32Z", "updated_at": "2024-02-15T18:51:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.10196v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T13:00:03.117157+00:00", "role": "primary", "sha256": "a1d4dc44a6572863df1d8cc68bff592c3a04a174a3dbc87ab5cdcd0d33be92b0", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/html/2402.10196v1"}], "id": "arxiv:2402.10196", "identifiers": {"arxiv": "2402.10196"}, "kind": "survey", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.10196"}], "name": "A Trembling House of Cards? Mapping Adversarial Attacks against Language Agents", "notes": "Adjacent conceptual security work: the paper describes hypothetical scenarios and a future generalist agent, rather than an implemented attack benchmark. Scope was checked against the retained abstract and full-text Sections 2.4 and 3; no results were independently reproduced.", "published_at": "2024-02-15", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Presents a conceptual perception–reasoning–action framework and 12 potential attack scenarios for language agents, including hypothetical web-navigation attacks.", "topics": ["safety-security", "survey", "robustness"], "updated_at": "2024-02-15", "url": "https://arxiv.org/abs/2402.10196"}
{"aliases": [], "authors": ["Wenkai Yang", "Xiaohan Bi", "Yankai Lin", "Sishuo Chen", "Jie Zhou", "Xu Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a00f68d32c29b7ab49b837363d41242db0bc2a4d2c4483a29eaf1d2c7aa88d2", "submitted_at": "2024-02-17T06:48:45Z", "updated_at": "2024-10-29T15:32:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.11208v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2402.11208", "identifiers": {"arxiv": "2402.11208"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.11208"}], "name": "Watch Out for Your Agents! Investigating Backdoor Threats to LLM-Based Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-29", "url": "https://arxiv.org/abs/2402.11208"}
{"aliases": [], "authors": ["Xinbei Ma", "Zhuosheng Zhang", "Hai Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "98930b10832d8169bdc60620f882b4b8f753440979c8e9a8de167b12d38d7060", "submitted_at": "2024-02-19T08:29:03Z", "updated_at": "2024-06-02T13:25:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.11941v3"}, "publication": {"authors": ["Xinbei Ma", "Zhuosheng Zhang", "Hai Zhao (赵海)"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.findings-acl.539", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-acl.539"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "9097-9110", "pdf_url": "https://aclanthology.org/2024.findings-acl.539.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "CoCo-Agent: A Comprehensive Cognitive MLLM Agent for Smartphone GUI Automation", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-acl.539", "venue": "Findings of the Association for Computational Linguistics: ACL 2024", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:49Z", "role": "primary", "sha256": "022e2ca49699ec713f11eec442b2139fc0ce4d075f0812f88844289749f13d6c", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-acl.539"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.11941", "identifiers": {"arxiv": "2402.11941"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.11941"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-acl.539"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-acl.539.pdf"}], "name": "CoCo-Agent: A Comprehensive Cognitive MLLM Agent for Smartphone GUI Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines screenshots, layout information, and action history with a factored action predictor for smartphone automation.", "topics": ["mobile-navigation", "screen-understanding"], "updated_at": "2024-06-02", "url": "https://arxiv.org/abs/2402.11941"}
{"aliases": [], "authors": ["Xueyang Feng", "Zhi-Yuan Chen", "Yujia Qin", "Yankai Lin", "Xu Chen", "Zhiyuan Liu", "Ji-Rong Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f7497bde3cd7602a4122005baf8a538a3997fc868e4e363b0b161e7261d9520e", "submitted_at": "2024-02-20T11:03:36Z", "updated_at": "2024-02-20T11:03:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.12914v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.12914", "identifiers": {"arxiv": "2402.12914"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.12914"}], "name": "Large Language Model-based Human-Agent Collaboration for Complex Task Solving", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-20", "url": "https://arxiv.org/abs/2402.12914"}
{"aliases": [], "authors": ["Qiao Jin", "Zhizheng Wang", "Yifan Yang", "Qingqing Zhu", "Donald Wright", "Thomas Huang", "W John Wilbur", "Zhe He", "Andrew Taylor", "Qingyu Chen", "Zhiyong Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b3f131aa0533caa0059707f2057050ab6f1127f43a52cede3771975bc2bb25f", "submitted_at": "2024-02-20T18:37:19Z", "updated_at": "2024-02-20T18:37:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.13225v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.13225", "identifiers": {"arxiv": "2402.13225"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.13225"}], "name": "AgentMD: Empowering Language Agents for Risk Prediction with Large-Scale Clinical Tool Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-20", "url": "https://arxiv.org/abs/2402.13225"}
{"aliases": [], "authors": ["Tianyu Gao", "Zirui Wang", "Adithya Bhaskar", "Danqi Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f3eed6e0021be438af9d056e97f71a838af905b948cfd6499423ab18df11e43", "submitted_at": "2024-02-21T19:01:03Z", "updated_at": "2024-02-21T19:01:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.14073v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.14073", "identifiers": {"arxiv": "2402.14073"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.14073"}], "name": "Improving Language Understanding from Screenshots", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-21", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Studies language understanding from text-rendered screenshots using joint image-patch and text prediction; a perception foundation relevant to screen-reading models.", "topics": ["screen-understanding"], "updated_at": "2024-02-21", "url": "https://arxiv.org/abs/2402.14073"}
{"aliases": [], "authors": ["Xiaogang Jia", "Denis Blessing", "Xinkai Jiang", "Moritz Reuss", "Atalay Donat", "Rudolf Lioutikov", "Gerhard Neumann"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf9425766ec985a88caa78939922b835dcda2673abb52057e459d1aabe56092e", "submitted_at": "2024-02-22T14:54:28Z", "updated_at": "2024-02-22T14:54:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.14606v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.14606", "identifiers": {"arxiv": "2402.14606"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.14606"}], "name": "Towards Diverse Behaviors: A Benchmark for Imitation Learning with Human Demonstrations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-22", "url": "https://arxiv.org/abs/2402.14606"}
{"aliases": [], "authors": ["Yang Deng", "Xuan Zhang", "Wenxuan Zhang", "Yifei Yuan", "See-Kiong Ng", "Tat-Seng Chua"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e4b233e423e9043b6d615c7b185a1eb0d47f646590e4fb9fd43ca156a1a1b16", "submitted_at": "2024-02-23T02:18:12Z", "updated_at": "2024-02-23T02:18:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.15057v1"}, "publication": {"authors": ["Yang Deng", "Xuan Zhang", "Wenxuan Zhang", "Yifei Yuan", "See Kiong Ng", "Tat-Seng Chua"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.477", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.477/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "8795-8812", "pdf_url": "https://aclanthology.org/2024.acl-long.477.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "On the Multi-turn Instruction Following for Conversational Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.477/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:49Z", "role": "primary", "sha256": "ba56d6d0a422b340672036d5c872cd00d54a0f12cfe1326a3be65e8702786381", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.acl-long.477/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2402.15057", "identifiers": {"arxiv": "2402.15057"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.15057"}, {"rel": "publication", "url": "https://aclanthology.org/2024.acl-long.477/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.acl-long.477.pdf"}], "name": "On the Multi-turn Instruction Following for Conversational Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "MT-Mind2Web evaluates web navigation through sequences of conversational user instructions. Self-MAP combines memory-augmented planning with reflection to preserve context and dependencies across those interaction turns.", "topics": ["evaluation", "memory", "self-reflection", "web-navigation"], "updated_at": "2024-02-23", "url": "https://arxiv.org/abs/2402.15057"}
{"aliases": [], "authors": ["Zhefan Wang", "Yuanqing Yu", "Wendi Zheng", "Weizhi Ma", "Min Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f8d3b182658e13c5d0df41623eff76fbe011db0b7291de3b56528e70f245266", "submitted_at": "2024-02-23T09:57:20Z", "updated_at": "2024-11-01T02:00:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.15235v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.15235", "identifiers": {"arxiv": "2402.15235"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.15235"}], "name": "MACRec: a Multi-Agent Collaboration Framework for Recommendation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-01", "url": "https://arxiv.org/abs/2402.15235"}
{"aliases": [], "authors": ["Jake Bruce", "Michael Dennis", "Ashley Edwards", "Jack Parker-Holder", "Yuge Shi", "Edward Hughes", "Matthew Lai", "Aditi Mavalankar", "Richie Steigerwald", "Chris Apps", "Yusuf Aytar", "Sarah Bechtle", "Feryal Behbahani", "Stephanie Chan", "Nicolas Heess", "Lucy Gonzalez", "Simon Osindero", "Sherjil Ozair", "Scott Reed", "Jingwei Zhang", "Konrad Zolna", "Jeff Clune", "Nando de Freitas", "Satinder Singh", "Tim Rocktäschel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68c2b5bbbb4a0897722d6e6325d3eca64ed34c5acb0c48406d8f64d3c6b17352", "submitted_at": "2024-02-23T15:47:26Z", "updated_at": "2024-02-23T15:47:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.15391v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.15391", "identifiers": {"arxiv": "2402.15391"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.15391"}], "name": "Genie: Generative Interactive Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-23", "url": "https://arxiv.org/abs/2402.15391"}
{"aliases": [], "authors": ["Jianguo Zhang", "Tian Lan", "Rithesh Murthy", "Zhiwei Liu", "Weiran Yao", "Ming Zhu", "Juntao Tan", "Thai Hoang", "Zuxin Liu", "Liangwei Yang", "Yihao Feng", "Shirley Kokane", "Tulika Awalgaonkar", "Juan Carlos Niebles", "Silvio Savarese", "Shelby Heinecke", "Huan Wang", "Caiming Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "252afc54882ef310b4d658876c5a185429bb85eca787a0b78f03d927627a6bc3", "submitted_at": "2024-02-23T18:56:26Z", "updated_at": "2024-11-09T00:28:26Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2402.15506v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2402.15506", "identifiers": {"arxiv": "2402.15506"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.15506"}], "name": "AgentOhana: Design Unified Data and Training Pipeline for Effective Agent Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-09", "url": "https://arxiv.org/abs/2402.15506"}
{"aliases": [], "authors": ["Fangzhou Wu", "Shutong Wu", "Yulong Cao", "Chaowei Xiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2cf1458c8c4134b2195394a8acba4919f9b3a5a7c79b1bea017ac4558123f69c", "submitted_at": "2024-02-26T19:01:54Z", "updated_at": "2024-02-26T19:01:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.16965v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2402.16965", "identifiers": {"arxiv": "2402.16965"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.16965"}], "name": "WIPI: A New Web Threat for LLM-Driven Web Agents", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-02-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies WIPI, a black-box indirect-instruction threat in which webpage content redirects LLM-driven web agents. Evaluation spans plugin-based agents, Web GPTs, and open-source web-agent implementations.", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2024-02-26", "url": "https://arxiv.org/abs/2402.16965"}
{"aliases": [], "authors": ["Rodrick Kuate Defo", "Steven L. Richardson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60a0ced1950b0cfb1d3b23915fd66322f31d47095ffaf334d4612256705e75be", "submitted_at": "2024-02-26T23:41:19Z", "updated_at": "2024-05-05T03:23:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.17083v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2402.17083", "identifiers": {"arxiv": "2402.17083"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.17083"}], "name": "Computing Using Shallow NV-Center Charges in Diamond", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-05", "url": "https://arxiv.org/abs/2402.17083"}
{"aliases": [], "authors": ["Nozomu Kobayashi", "Yoshiyuki Suimon", "Koichi Miyamoto"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8f97eb8c47e32209d5782fe273619e2b6ea68824827cb6e750b7d4ac19418a7", "submitted_at": "2024-02-27T02:29:24Z", "updated_at": "2026-04-09T00:31:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.17148v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2402.17148", "identifiers": {"arxiv": "2402.17148"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.17148"}], "name": "Time series generation for option pricing on quantum computers using tensor network", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-09", "url": "https://arxiv.org/abs/2402.17148"}
{"aliases": [], "authors": ["Raghav Kapoor", "Yash Parag Butala", "Melisa Russak", "Jing Yu Koh", "Kiran Kamble", "Waseem Alshikh", "Ruslan Salakhutdinov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54b548ac2bee529d3a769e7ea14b6bfce2270cdf349d832cc8c146e7d779b81a", "submitted_at": "2024-02-27T14:47:53Z", "updated_at": "2024-07-21T23:16:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.17553v3"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "id": "arxiv:2402.17553", "identifiers": {"arxiv": "2402.17553"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.17553"}], "name": "OmniACT: A Dataset and Benchmark for Enabling Multimodal Generalist Autonomous Agents for Desktop and Web", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-02-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a dataset and benchmark for generating executable task scripts from screenshots and visually grounded desktop or web instructions.", "topics": ["evaluation", "grounding"], "updated_at": "2024-07-21", "url": "https://arxiv.org/abs/2402.17553"}
{"aliases": [], "authors": ["Marco Bellagente", "Jonathan Tow", "Dakota Mahan", "Duy Phung", "Maksym Zhuravinskyi", "Reshinth Adithyan", "James Baicoianu", "Ben Brooks", "Nathan Cooper", "Ashish Datta", "Meng Lee", "Emad Mostaque", "Michael Pieler", "Nikhil Pinnaparju", "Paulo Rocha", "Harry Saini", "Hannah Teufel", "Niccolo Zanichelli", "Carlos Riquelme"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "989276fd7ba547e33e32522b0fc50fc1ed81c2bdde7d69f10bf55b5a4dd0ad68", "submitted_at": "2024-02-27T19:00:07Z", "updated_at": "2024-02-27T19:00:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.17834v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.17834", "identifiers": {"arxiv": "2402.17834"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.17834"}], "name": "Stable LM 2 1.6B Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-27", "url": "https://arxiv.org/abs/2402.17834"}
{"aliases": [], "authors": ["Corby Rosset", "Ho-Lam Chung", "Guanghui Qin", "Ethan C. Chau", "Zhuo Feng", "Ahmed Awadallah", "Jennifer Neville", "Nikhil Rao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1b538ae272b6b534fc8697b9a43078abd84b755eec2d666d221bd87c6f91411e", "submitted_at": "2024-02-27T21:27:16Z", "updated_at": "2024-02-27T21:27:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.17896v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2402.17896", "identifiers": {"arxiv": "2402.17896"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.17896"}], "name": "Researchy Questions: A Dataset of Multi-Perspective, Decompositional Questions for LLM Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-27", "url": "https://arxiv.org/abs/2402.17896"}
{"aliases": [], "authors": ["Tan Zhi-Xuan", "Lance Ying", "Vikash Mansinghka", "Joshua B. Tenenbaum"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8d3617bfd5366e3ce9cf0fb51cee0a0dcc31c8a7f26b7adde68a3b7d8a737c4b", "submitted_at": "2024-02-27T23:06:53Z", "updated_at": "2024-02-27T23:06:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.17930v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.17930", "identifiers": {"arxiv": "2402.17930"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.17930"}], "name": "Pragmatic Instruction Following and Goal Assistance via Cooperative Language-Guided Inverse Planning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-27", "url": "https://arxiv.org/abs/2402.17930"}
{"aliases": [], "authors": ["George Karabatsos"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a6589e8eed68504eef005067d0636cdec18449a4c972ae1fad3f73da2c747858", "submitted_at": "2024-02-28T16:21:20Z", "updated_at": "2024-09-11T17:44:35Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2402.18450v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2402.18450", "identifiers": {"arxiv": "2402.18450"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.18450"}], "name": "Copula Approximate Bayesian Computation Using Distribution Random Forests", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-11", "url": "https://arxiv.org/abs/2402.18450"}
{"aliases": [], "authors": ["Takashi Imoto", "Yuki Susa", "Ryoji Miyazaki", "Yuichiro Matsuzaki"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5cf10b4e648cbcefbba49875190bd229b66c8ee626916558df74de113b712f02", "submitted_at": "2024-02-29T12:47:29Z", "updated_at": "2025-03-12T04:29:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.19114v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2402.19114", "identifiers": {"arxiv": "2402.19114"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.19114"}], "name": "Universal quantum computation using quantum annealing with the transverse-field Ising Hamiltonian", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-12", "url": "https://arxiv.org/abs/2402.19114"}
{"aliases": [], "authors": ["Yifei Zhou", "Andrea Zanette", "Jiayi Pan", "Sergey Levine", "Aviral Kumar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f64b4dff07882ddff979bbbe7acc9a06f0637568650f6ccba37fe4b12e6e178f", "submitted_at": "2024-02-29T18:45:56Z", "updated_at": "2024-02-29T18:45:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.19446v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2402.19446", "identifiers": {"arxiv": "2402.19446"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2402.19446"}], "name": "ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-02-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-02-29", "url": "https://arxiv.org/abs/2402.19446"}
{"aliases": [], "authors": ["A. H. Abbas", "Ivan S. Maksymov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d1411cdfbb5cbca1c4c9b309d5db448f75747033c54bf41fac98d2c093197123", "submitted_at": "2024-03-01T22:59:41Z", "updated_at": "2024-03-01T22:59:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.01024v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2403.01024", "identifiers": {"arxiv": "2403.01024"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.01024"}], "name": "Reservoir Computing Using Measurement-Controlled Quantum Dynamics", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-01", "url": "https://arxiv.org/abs/2403.01024"}
{"aliases": [], "authors": ["Ziniu Hu", "Ahmet Iscen", "Aashi Jain", "Thomas Kipf", "Yisong Yue", "David A. Ross", "Cordelia Schmid", "Alireza Fathi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "936e6f8054da889481c76003e8c22e91b9896df6625e274ccac300129d6d5bff", "submitted_at": "2024-03-02T16:16:26Z", "updated_at": "2024-03-02T16:16:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.01248v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.01248", "identifiers": {"arxiv": "2403.01248"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.01248"}], "name": "SceneCraft: An LLM Agent for Synthesizing 3D Scene as Blender Code", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-02", "url": "https://arxiv.org/abs/2403.01248"}
{"aliases": [], "authors": ["Grischa Beneke", "Thomas Brian Winkler", "Klaus Raab", "Maarten A. Brems", "Fabian Kammerbauer", "Pascal Gerhards", "Klaus Knobloch", "Johan Mentink", "Mathias Kläui"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ed1e985985969c2d8d87d0e3de9e56c14b84b4269bda3b2ff8c89240698f9ba2", "submitted_at": "2024-03-04T09:32:13Z", "updated_at": "2024-03-04T09:32:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.01877v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2403.01877", "identifiers": {"arxiv": "2403.01877"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.01877"}], "name": "Gesture recognition with Brownian reservoir computing using geometrically confined skyrmion dynamics", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-04", "url": "https://arxiv.org/abs/2403.01877"}
{"aliases": [], "authors": ["Yifan Song", "Da Yin", "Xiang Yue", "Jie Huang", "Sujian Li", "Bill Yuchen Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af4dbaf0179c52ad858c63b925d4c798ebf5b797a15f0b0d60c3690bdb8ca22a", "submitted_at": "2024-03-04T21:50:29Z", "updated_at": "2024-07-10T17:36:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.02502v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2403.02502", "identifiers": {"arxiv": "2403.02502"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.02502"}], "name": "Trial and Error: Exploration-Based Trajectory Optimization for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-10", "url": "https://arxiv.org/abs/2403.02502"}
{"aliases": [], "authors": ["Jiwen Zhang", "Jihao Wu", "Yihua Teng", "Minghui Liao", "Nuo Xu", "Xiao Xiao", "Zhongyu Wei", "Duyu Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4de9e7d68aa8450cab65ecb300524e85b1d2b2a91d58fecef370deceaba64b1e", "submitted_at": "2024-03-05T07:09:35Z", "updated_at": "2024-07-13T02:12:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.02713v2"}, "publication": {"authors": ["Jiwen Zhang (张霁雯)", "Jihao Wu", "Teng Yihua", "Minghui Liao", "Nuo Xu", "Xiao Xiao", "Zhongyu Wei (魏忠钰)", "Duyu Tang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.702", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.702/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "12016-12031", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.702.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Android in the Zoo: Chain-of-Action-Thought for GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.702/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:50Z", "role": "primary", "sha256": "87a46c27e83152e7741c2398f790320f937b8fd770fa7ac7d3a68395427e87d2", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-emnlp.702/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2403.02713", "identifiers": {"arxiv": "2403.02713"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.02713"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-emnlp.702/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-emnlp.702.pdf"}], "name": "Android in the Zoo: Chain-of-Action-Thought for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-03-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Chain-of-Action-Thought and Android-In-The-Zoo annotations that connect prior actions, screen observations, intended actions, and expected outcomes.", "topics": ["mobile-navigation", "planning"], "updated_at": "2024-07-13", "url": "https://arxiv.org/abs/2403.02713"}
{"aliases": [], "authors": ["Bo Wang", "Tianxiang Sun", "Hang Yan", "Siyin Wang", "Qingyuan Cheng", "Xipeng Qiu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86bca1cf8f8ee8dee6ed0ec94c2b6a838d84204451dc2f2ff4f20eb41c57e237", "submitted_at": "2024-03-05T08:25:11Z", "updated_at": "2024-03-05T08:25:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.02757v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.02757", "identifiers": {"arxiv": "2403.02757"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.02757"}], "name": "In-Memory Learning: A Declarative Learning Framework for Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-05", "url": "https://arxiv.org/abs/2403.02757"}
{"aliases": [], "authors": ["Weihao Tan", "Wentao Zhang", "Xinrun Xu", "Haochong Xia", "Ziluo Ding", "Boyu Li", "Bohan Zhou", "Junpeng Yue", "Jiechuan Jiang", "Yewen Li", "Ruyi An", "Molei Qin", "Chuqiao Zong", "Longtao Zheng", "Yujie Wu", "Xiaoqiang Chai", "Yifei Bi", "Tianbao Xie", "Pengjie Gu", "Xiyun Li", "Ceyao Zhang", "Long Tian", "Chaojie Wang", "Xinrun Wang", "Börje F. Karlsson", "Bo An", "Shuicheng Yan", "Zongqing Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63c68c596f1c81485e77f98beae696deca9785a5d8725ab4456ee35870e64a99", "submitted_at": "2024-03-05T18:22:29Z", "updated_at": "2024-07-02T17:23:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.03186v3"}, "publication": {"authors": ["Weihao Tan", "Wentao Zhang", "Xinrun Xu", "Haochong Xia", "Ziluo Ding", "Boyu Li", "Bohan Zhou", "Junpeng Yue", "Jiechuan Jiang", "Yewen Li", "Ruyi An", "Molei Qin", "Chuqiao Zong", "Longtao Zheng", "Yujie Wu", "Xiaoqiang Chai", "Yifei Bi", "Tianbao Xie", "Pengjie Gu", "Xiyun Li", "Ceyao Zhang", "Long Tian", "Chaojie Wang", "Xinrun Wang", "Börje F. Karlsson", "Bo An", "Shuicheng Yan", "Zongqing Lu"], "date_basis": "Explicit publisher metadata: 2025/10/06; precision=day.", "date_parts": [2025, 10, 6], "doi": null, "event_dates": "13--19 Jul 2025", "evidence_urls": ["https://proceedings.mlr.press/v267/tan25h.html"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "58658-58725", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/tan25h/tan25h.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "Cradle: Empowering Foundation Agents towards General Computer Control", "type": "conference-paper", "url": "https://proceedings.mlr.press/v267/tan25h.html", "venue": "Proceedings of the 42nd International Conference on Machine Learning", "volume": "267"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:47+00:00", "role": "primary", "sha256": "c081bdd38f20d6c3aac756da474b8cb8a6d782ece213f67ebc333bf2b99de2fc", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v267/tan25h.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2403.03186", "identifiers": {"arxiv": "2403.03186"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.03186"}, {"rel": "publication", "url": "https://proceedings.mlr.press/v267/tan25h.html"}, {"rel": "publication_pdf", "url": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/tan25h/tan25h.pdf"}], "name": "Cradle: Empowering Foundation Agents Towards General Computer Control", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-03-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses screenshot observations, planning modules, and generated low-level keyboard and mouse control for general computer interaction.", "topics": ["planning", "screen-understanding"], "updated_at": "2024-07-02", "url": "https://arxiv.org/abs/2403.03186"}
{"aliases": [], "authors": ["Yibin Chen", "Yifu Yuan", "Zeyu Zhang", "Yan Zheng", "Jinyi Liu", "Fei Ni", "Jianye Hao", "Hangyu Mao", "Fuzheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f1fde01350fea1db3d607c19efee666875d94a0d057069266fdf2b46293ada28", "submitted_at": "2024-03-06T11:48:08Z", "updated_at": "2025-03-03T06:56:29Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.03636v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.03636", "identifiers": {"arxiv": "2403.03636"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.03636"}], "name": "SheetAgent: Towards A Generalist Agent for Spreadsheet Reasoning and Manipulation via Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-03", "url": "https://arxiv.org/abs/2403.03636"}
{"aliases": [], "authors": ["Zekai Zhang", "Yiduo Guo", "Yaobo Liang", "Dongyan Zhao", "Nan Duan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51abf6b165b84c647c9a4870a6441900cac5e9a933199fe8a47d1b7f065eb11e", "submitted_at": "2024-03-06T15:33:32Z", "updated_at": "2024-03-06T15:33:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.03788v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2403.03788", "identifiers": {"arxiv": "2403.03788"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.03788"}], "name": "PPTC-R benchmark: Towards Evaluating the Robustness of Large Language Models for PowerPoint Task Completion", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The benchmark directly concerns PowerPoint task automation but executes through APIs, so it remains adjacent.", "published_at": "2024-03-06", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Benchmarks robustness of API-based PowerPoint completion under sentence-level, semantic, and multilingual instruction perturbations and changes to the available API set.", "topics": ["evaluation", "robustness", "workflow"], "updated_at": "2024-03-06", "url": "https://arxiv.org/abs/2403.03788"}
{"aliases": [], "authors": ["Junsong Chen", "Chongjian Ge", "Enze Xie", "Yue Wu", "Lewei Yao", "Xiaozhe Ren", "Zhongdao Wang", "Ping Luo", "Huchuan Lu", "Zhenguo Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c73a276d0b06f927e39a240cbae27683622e531476a7cbbc1a88c0ffc2c931a", "submitted_at": "2024-03-07T17:41:37Z", "updated_at": "2024-03-17T16:59:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.04692v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.04692", "identifiers": {"arxiv": "2403.04692"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.04692"}], "name": "PixArt-Σ: Weak-to-Strong Training of Diffusion Transformer for 4K Text-to-Image Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-17", "url": "https://arxiv.org/abs/2403.04692"}
{"aliases": [], "authors": ["Xiaogeng Liu", "Zhiyuan Yu", "Yizhe Zhang", "Ning Zhang", "Chaowei Xiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b74bfce807087c59db9e9b030ec83e5c9240c13e6cc9470a3623b66520c76745", "submitted_at": "2024-03-07T23:46:20Z", "updated_at": "2024-03-07T23:46:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.04957v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.04957", "identifiers": {"arxiv": "2403.04957"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.04957"}], "name": "Automatic and Universal Prompt Injection Attacks against Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-07", "url": "https://arxiv.org/abs/2403.04957"}
{"aliases": [], "authors": ["Shuhong Xiao", "Yunnong Chen", "Yaxuan Song", "Liuqing Chen", "Lingyun Sun", "Yankun Zhen", "Yanfang Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f4bfed18cba0c946a8ae013c819d84d7f7b2c639b2e3471d6bf7890248c4530b", "submitted_at": "2024-03-08T01:52:44Z", "updated_at": "2024-03-08T01:52:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.04984v1"}}, "environments": ["android", "ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2403.04984", "identifiers": {"arxiv": "2403.04984"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.04984"}], "name": "UI Semantic Group Detection: Grouping UI Elements with Similar Semantics in Mobile Graphical User Interface", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2024-03-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "UISCGD detects semantic groups of adjacent text and non-text UI elements. It learns from Android and iOS screenshots to support interface retrieval, UI-to-code structure and screen-reader accessibility.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-03-08", "url": "https://arxiv.org/abs/2403.04984"}
{"aliases": [], "authors": ["Omer Rathore", "Alastair Basden", "Nicholas Chancellor", "Halim Kusumaatmaja"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4e5cfc71460af6ecc5b7c54facfdc1fc3cd86c79229f5a56e03fdb405139ebb0", "submitted_at": "2024-03-08T12:58:12Z", "updated_at": "2024-03-08T12:58:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.05278v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2403.05278", "identifiers": {"arxiv": "2403.05278"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.05278"}], "name": "Load Balancing For High Performance Computing Using Quantum Annealing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-08", "url": "https://arxiv.org/abs/2403.05278"}
{"aliases": [], "authors": ["Alexandre Drouin", "Maxime Gasse", "Massimo Caccia", "Issam H. Laradji", "Manuel Del Verme", "Tom Marty", "Léo Boisvert", "Megh Thakkar", "Quentin Cappart", "David Vazquez", "Nicolas Chapados", "Alexandre Lacoste"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8033f9c26a5f95a49000374a6edebb4dead83e100ebcf5226007f06b7573c305", "submitted_at": "2024-03-12T14:58:45Z", "updated_at": "2024-07-23T06:19:28Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2403.07718v5"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2403.07718", "identifiers": {"arxiv": "2403.07718"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.07718"}], "name": "WorkArena: How Capable Are Web Agents at Solving Common Knowledge Work Tasks?", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-03-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces enterprise web tasks on ServiceNow and BrowserGym interfaces for evaluating browser-based agents.", "topics": ["evaluation", "workflow"], "updated_at": "2024-07-23", "url": "https://arxiv.org/abs/2403.07718"}
{"aliases": [], "authors": ["Gemma Team", "Thomas Mesnard", "Cassidy Hardin", "Robert Dadashi", "Surya Bhupatiraju", "Shreya Pathak", "Laurent Sifre", "Morgane Rivière", "Mihir Sanjay Kale", "Juliette Love", "Pouya Tafti", "Léonard Hussenot", "Pier Giuseppe Sessa", "Aakanksha Chowdhery", "Adam Roberts", "Aditya Barua", "Alex Botev", "Alex Castro-Ros", "Ambrose Slone", "Amélie Héliou", "Andrea Tacchetti", "Anna Bulanova", "Antonia Paterson", "Beth Tsai", "Bobak Shahriari", "Charline Le Lan", "Christopher A. Choquette-Choo", "Clément Crepy", "Daniel Cer", "Daphne Ippolito", "David Reid", "Elena Buchatskaya", "Eric Ni", "Eric Noland", "Geng Yan", "George Tucker", "George-Christian Muraru", "Grigory Rozhdestvenskiy", "Henryk Michalewski", "Ian Tenney", "Ivan Grishchenko", "Jacob Austin", "James Keeling", "Jane Labanowski", "Jean-Baptiste Lespiau", "Jeff Stanway", "Jenny Brennan", "Jeremy Chen", "Johan Ferret", "Justin Chiu", "Justin Mao-Jones", "Katherine Lee", "Kathy Yu", "Katie Millican", "Lars Lowe Sjoesund", "Lisa Lee", "Lucas Dixon", "Machel Reid", "Maciej Mikuła", "Mateo Wirth", "Michael Sharman", "Nikolai Chinaev", "Nithum Thain", "Olivier Bachem", "Oscar Chang", "Oscar Wahltinez", "Paige Bailey", "Paul Michel", "Petko Yotov", "Rahma Chaabouni", "Ramona Comanescu", "Reena Jana", "Rohan Anil", "Ross McIlroy", "Ruibo Liu", "Ryan Mullins", "Samuel L Smith", "Sebastian Borgeaud", "Sertan Girgin", "Sholto Douglas", "Shree Pandya", "Siamak Shakeri", "Soham De", "Ted Klimenko", "Tom Hennigan", "Vlad Feinberg", "Wojciech Stokowiec", "Yu-hui Chen", "Zafarali Ahmed", "Zhitao Gong", "Tris Warkentin", "Ludovic Peran", "Minh Giang", "Clément Farabet", "Oriol Vinyals", "Jeff Dean", "Koray Kavukcuoglu", "Demis Hassabis", "Zoubin Ghahramani", "Douglas Eck", "Joelle Barral", "Fernando Pereira", "Eli Collins", "Armand Joulin", "Noah Fiedel", "Evan Senter", "Alek Andreev", "Kathleen Kenealy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0dca710f82058729257a49070b14af85da99e6f6a6c70457179d48df6d28704", "submitted_at": "2024-03-13T06:59:16Z", "updated_at": "2024-04-16T12:52:47Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2403.08295v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.08295", "identifiers": {"arxiv": "2403.08295"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.08295"}], "name": "Gemma: Open Models Based on Gemini Research and Technology", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-16", "url": "https://arxiv.org/abs/2403.08295"}
{"aliases": [], "authors": ["Andrés Muñoz", "Daniel Borrajo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aacb4d2cac31e478aa56ee25b6b39e5ba8fe164f89102e665f8fb25cf3822e7b", "submitted_at": "2024-03-15T10:26:52Z", "updated_at": "2024-08-27T18:36:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.10170v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2403.10170", "identifiers": {"arxiv": "2403.10170"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.10170"}], "name": "Computer User Interface Understanding. A New Dataset and a Learning Framework", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-03-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces desktop interaction videos and a framework combining synthetic examples with hierarchical contrastive learning for interface classification.", "topics": ["screen-understanding", "video", "data-generation"], "updated_at": "2024-08-27", "url": "https://arxiv.org/abs/2403.10170"}
{"aliases": [], "authors": ["Xinrun Xu", "Yuxin Wang", "Chaoyi Xu", "Ziluo Ding", "Jiechuan Jiang", "Zhiming Ding", "Börje F. Karlsson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8497e90903223242fa595c853eec6eada92dc8833239d8e92f05517802712735", "submitted_at": "2024-03-15T12:37:12Z", "updated_at": "2024-03-15T12:37:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.10249v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2403.10249", "identifiers": {"arxiv": "2403.10249"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.10249"}], "name": "A Survey on Game Playing Agents and Large Models: Methods, Applications, and Challenges", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-15", "url": "https://arxiv.org/abs/2403.10249"}
{"aliases": [], "authors": ["Kevin Xu", "Yeganeh Kordi", "Tanay Nayak", "Adi Asija", "Yizhong Wang", "Kate Sanders", "Adam Byerly", "Jingyu Zhang", "Benjamin Van Durme", "Daniel Khashabi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d82424b84ebddfeb005ffae38c130c6a9c7dae7adcdc59e2b7d001284b785573", "submitted_at": "2024-03-18T16:06:30Z", "updated_at": "2025-02-22T01:33:54Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2403.11905v4"}, "publication": {"authors": ["Kevin Xu", "Yeganeh Kordi", "Tanay Nayak", "Adi Asija", "Yizhong Wang", "Kate Sanders", "Adam Byerly", "Jingyu Zhang", "Benjamin Van Durme", "Daniel Khashabi"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-long.188", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-long.188/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "3694-3710", "pdf_url": "https://aclanthology.org/2025.naacl-long.188.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "TurkingBench: A Challenge Benchmark for Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-long.188/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:50Z", "role": "primary", "sha256": "56537fe44b4655949e8c70cc49a4ed7304a8201a12f75ddb3caa5df92d6f8e16", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-long.188/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2403.11905", "identifiers": {"arxiv": "2403.11905"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.11905"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-long.188/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-long.188.pdf"}], "name": "Tur[k]ingBench: A Challenge Benchmark for Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-03-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates web agents on multimodal annotation tasks adapted from HTML interfaces originally built for crowdsourcing workers.", "topics": ["evaluation", "web-navigation"], "updated_at": "2025-02-22", "url": "https://arxiv.org/abs/2403.11905"}
{"aliases": [], "authors": ["Enshen Zhou", "Yiran Qin", "Zhenfei Yin", "Yuzhou Huang", "Ruimao Zhang", "Lu Sheng", "Yu Qiao", "Jing Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "973133be9a1bc8ed5d6d1d9c6efb554d9114b1f16381456f7966b314ff8b1133", "submitted_at": "2024-03-18T17:59:42Z", "updated_at": "2024-03-19T14:52:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.12037v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2403.12037", "identifiers": {"arxiv": "2403.12037"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.12037"}], "name": "MineDreamer: Learning to Follow Instructions via Chain-of-Imagination for Simulated-World Control", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-19", "url": "https://arxiv.org/abs/2403.12037"}
{"aliases": [], "authors": ["Zhuoshi Pan", "Qianhui Wu", "Huiqiang Jiang", "Menglin Xia", "Xufang Luo", "Jue Zhang", "Qingwei Lin", "Victor Rühle", "Yuqing Yang", "Chin-Yew Lin", "H. Vicky Zhao", "Lili Qiu", "Dongmei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b524260acfb004218ae0c5b1c649931c5f1f75444d0a8d37b72b655b8789e9d9", "submitted_at": "2024-03-19T17:59:56Z", "updated_at": "2024-08-12T04:48:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.12968v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.12968", "identifiers": {"arxiv": "2403.12968"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.12968"}], "name": "LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-12", "url": "https://arxiv.org/abs/2403.12968"}
{"aliases": [], "authors": ["Charles Goddard", "Shamane Siriwardhana", "Malikeh Ehghaghi", "Luke Meyers", "Vlad Karpukhin", "Brian Benedict", "Mark McQuade", "Jacob Solawetz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1e171adfc1aeca87e996beed942c92bd24b7ce8e822d30327eb88358b2c3d638", "submitted_at": "2024-03-20T02:38:01Z", "updated_at": "2025-01-09T22:21:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.13257v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.13257", "identifiers": {"arxiv": "2403.13257"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.13257"}], "name": "Arcee's MergeKit: A Toolkit for Merging Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-09", "url": "https://arxiv.org/abs/2403.13257"}
{"aliases": [], "authors": ["Keegan Hines", "Gary Lopez", "Matthew Hall", "Federico Zarfati", "Yonatan Zunger", "Emre Kiciman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "99e4ff426cbfdb6d73957ebf1922ec8d3869ae59dcb75cd3417f18e51c87b2d9", "submitted_at": "2024-03-20T15:26:23Z", "updated_at": "2024-03-20T15:26:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.14720v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.14720", "identifiers": {"arxiv": "2403.14720"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.14720"}], "name": "Defending Against Indirect Prompt Injection Attacks With Spotlighting", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-20", "url": "https://arxiv.org/abs/2403.14720"}
{"aliases": [], "authors": ["Seongjun Jeong", "Gi-Cheon Kang", "Seongho Choi", "Joochan Kim", "Byoung-Tak Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7a059a1dc2c123204ab92adb811890099a910a52348cb5d6fa858c4106d6bc3e", "submitted_at": "2024-03-22T09:15:36Z", "updated_at": "2025-10-31T01:59:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.15049v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.15049", "identifiers": {"arxiv": "2403.15049"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.15049"}], "name": "Continual Vision-and-Language Navigation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-31", "url": "https://arxiv.org/abs/2403.15049"}
{"aliases": [], "authors": ["Nick Heppert", "Max Argus", "Tim Welschehold", "Thomas Brox", "Abhinav Valada"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d475940aabcedc2a2c4b7a86330ee399abafd39f56f7a49d3c47df34150b73e", "submitted_at": "2024-03-22T13:46:51Z", "updated_at": "2024-10-02T11:41:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.15203v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2403.15203", "identifiers": {"arxiv": "2403.15203"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.15203"}], "name": "DITTO: Demonstration Imitation by Trajectory Transformation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-02", "url": "https://arxiv.org/abs/2403.15203"}
{"aliases": [], "authors": ["Kai Mei", "Xi Zhu", "Wujiang Xu", "Wenyue Hua", "Mingyu Jin", "Zelong Li", "Shuyuan Xu", "Ruosong Ye", "Yingqiang Ge", "Yongfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "351f10b771f554317b058b00b6b406cc947656a5d915af04ea69e2e30827a79a", "submitted_at": "2024-03-25T17:32:23Z", "updated_at": "2025-08-12T14:37:01Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2403.16971v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2403.16971", "identifiers": {"arxiv": "2403.16971"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.16971"}], "name": "AIOS: LLM Agent Operating System", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-12", "url": "https://arxiv.org/abs/2403.16971"}
{"aliases": [], "authors": ["Longtao Zheng", "Zhiyuan Huang", "Zhenghai Xue", "Xinrun Wang", "Bo An", "Shuicheng Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe1432b46c15566182418bf3fa2f918607e8b0e9a383459416d03e28dc7e347c", "submitted_at": "2024-03-26T17:54:15Z", "updated_at": "2025-02-14T08:13:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.17918v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2403.17918", "identifiers": {"arxiv": "2403.17918"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.17918"}], "name": "AgentStudio: A Toolkit for Building General Virtual Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-03-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "AgentStudio combines an interactive GUI/API environment with tools for creating tasks, annotating interface elements, and labeling actions in videos. Its online suite and GroundUI, IDMBench, and CriticBench datasets examine interaction, grounding, learning from video, and success detection.", "topics": ["evaluation", "sandbox", "grounding", "video"], "updated_at": "2025-02-14", "url": "https://arxiv.org/abs/2403.17918"}
{"aliases": [], "authors": ["Weifeng Lin", "Xinyu Wei", "Ruichuan An", "Peng Gao", "Bocheng Zou", "Yulin Luo", "Siyuan Huang", "Shanghang Zhang", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c6636a8203076662b5d06ffd29c6578d68b8a79a11d9f3b58483fc2dbbf2c96b", "submitted_at": "2024-03-29T16:26:20Z", "updated_at": "2025-02-22T14:02:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.20271v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2403.20271", "identifiers": {"arxiv": "2403.20271"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2403.20271"}], "name": "Draw-and-Understand: Leveraging Visual Prompts to Enable MLLMs to Comprehend What You Want", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-22", "url": "https://arxiv.org/abs/2403.20271"}
{"aliases": [], "authors": ["Joel Ruben Antony Moniz", "Soundarya Krishnan", "Melis Ozyildirim", "Prathamesh Saraf", "Halim Cagri Ates", "Yuan Zhang", "Hong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a3573a5cf66737178f164d9809b3e6dcadb7c6b2b4ae7d95283e8a136bf78851", "submitted_at": "2024-03-29T17:59:06Z", "updated_at": "2024-08-19T03:06:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.20329v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2403.20329", "identifiers": {"arxiv": "2403.20329"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.20329"}], "name": "ReALM: Reference Resolution As Language Modeling", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. This is reference resolution involving screen entities, not an evaluated end-to-end GUI controller.", "published_at": "2024-03-29", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Frames conversational and on-screen reference resolution as language modeling, allowing textual representations of screen entities to support user-reference interpretation. Evaluation separately examines on-screen references against an existing resolver and language-model baselines.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-08-19", "url": "https://arxiv.org/abs/2403.20329"}
{"aliases": [], "authors": ["Ashish Rana", "Michael Oesterle", "Jannik Brinkmann"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ef15ce18975ec564c55993465964dca9a4cbe3e087d52e69a4ba84a5ecf465b", "submitted_at": "2024-04-01T14:19:00Z", "updated_at": "2024-04-14T19:54:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.01131v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2404.01131", "identifiers": {"arxiv": "2404.01131"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.01131"}], "name": "GOV-REK: Governed Reward Engineering Kernels for Designing Robust Multi-Agent Reinforcement Learning Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-14", "url": "https://arxiv.org/abs/2404.01131"}
{"aliases": [], "authors": ["Nikhil Pinnaparaju", "Reshinth Adithyan", "Duy Phung", "Jonathan Tow", "James Baicoianu", "Ashish Datta", "Maksym Zhuravinskyi", "Dakota Mahan", "Marco Bellagente", "Carlos Riquelme", "Nathan Cooper"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a093ee5a9d9821e2b0118032715e040d5a4a9071cb2cc445a4c819a71a4d3450", "submitted_at": "2024-04-01T16:39:36Z", "updated_at": "2024-04-01T16:39:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.01226v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2404.01226", "identifiers": {"arxiv": "2404.01226"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.01226"}], "name": "Stable Code Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-01", "url": "https://arxiv.org/abs/2404.01226"}
{"aliases": [], "authors": ["Patrick Chao", "Edoardo Debenedetti", "Alexander Robey", "Maksym Andriushchenko", "Francesco Croce", "Vikash Sehwag", "Edgar Dobriban", "Nicolas Flammarion", "George J. Pappas", "Florian Tramer", "Hamed Hassani", "Eric Wong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52a15c2b3dcfbabdef2cb7bccedcd7cb22fffe2d99acbae418d2a28f4cc24b1c", "submitted_at": "2024-03-28T02:44:02Z", "updated_at": "2024-10-31T22:26:40Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2404.01318v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2404.01318", "identifiers": {"arxiv": "2404.01318"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.01318"}], "name": "JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-31", "url": "https://arxiv.org/abs/2404.01318"}
{"aliases": [], "authors": ["Ashot Avanesov", "Alexander Shurinov", "Ivan Dyakonov", "Stanislav Straupe"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c9f02236d664249f5cef48925603e3db6672d74e0d4a7901201aa85d1060cf0", "submitted_at": "2024-04-01T20:58:35Z", "updated_at": "2025-05-14T04:07:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.01477v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2404.01477", "identifiers": {"arxiv": "2404.01477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.01477"}], "name": "Building a fusion-based quantum computer using teleported gates", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-14", "url": "https://arxiv.org/abs/2404.01477"}
{"aliases": [], "authors": ["Wei Chen", "Zhiyuan Li", "Mingyuan Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c2d8230ae5e7972b29d699e4960a01dbd4d9d920c74ba6e76d270a6f196eadc0", "submitted_at": "2024-04-02T01:29:28Z", "updated_at": "2026-07-20T02:25:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.01549v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.01549", "identifiers": {"arxiv": "2404.01549"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.01549"}], "name": "Octopus: On-device language model for function calling of software APIs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-20", "url": "https://arxiv.org/abs/2404.01549"}
{"aliases": [], "authors": ["Wei Chen", "Zhiyuan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5bb33264d11af4c0d8763c2f672a485b0f9699131636695bc4e14cd207b42ad", "submitted_at": "2024-04-02T09:01:32Z", "updated_at": "2026-09-11T20:45:11Z", "version": 7, "versioned_url": "https://arxiv.org/abs/2404.01744v7"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.01744", "identifiers": {"arxiv": "2404.01744"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.01744"}], "name": "Octopus v2: On-device language model for super agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-11", "url": "https://arxiv.org/abs/2404.01744"}
{"aliases": [], "authors": ["Tian Huang", "Chun Yu", "Weinan Shi", "Zijian Peng", "David Yang", "Weiqi Sun", "Yuanchun Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b2f357e166bddc2bbb5db4ec102721fc12fd27ba45d953193be0354fa5c7bce", "submitted_at": "2024-04-03T05:32:05Z", "updated_at": "2025-03-18T09:59:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.02475v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.02475", "identifiers": {"arxiv": "2404.02475"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.02475"}], "name": "Prompt2Task: Automating UI Tasks on Smartphones from Textual Prompts", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Prompt2Task converts textual goals or procedures into smartphone UI automation using agents for intent interpretation, external-information handling, and execution. User feedback and accumulated task knowledge support subsequent task generation and operation.", "topics": ["mobile-navigation", "human-in-the-loop", "workflow"], "updated_at": "2025-03-18", "url": "https://arxiv.org/abs/2404.02475"}
{"aliases": [], "authors": ["Hanyu Lai", "Xiao Liu", "Iat Long Iong", "Shuntian Yao", "Yuxuan Chen", "Pengbo Shen", "Hao Yu", "Hanchen Zhang", "Xiaohan Zhang", "Yuxiao Dong", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "340206da25df688ff73da617fd66a872586d1e36981c431d18331cee9b9669e0", "submitted_at": "2024-04-04T17:58:40Z", "updated_at": "2024-10-12T05:05:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.03648v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.03648", "identifiers": {"arxiv": "2404.03648"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.03648"}], "name": "AutoWebGLM: A Large Language Model-based Web Navigating Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds a web agent through simplified HTML observations, human-and-AI browsing data, curriculum training, and reinforcement-learning refinement.", "topics": ["web-navigation", "reinforcement-learning", "data-generation"], "updated_at": "2024-10-12", "url": "https://arxiv.org/abs/2404.03648"}
{"aliases": [], "authors": ["Yuntong Zhang", "Haifeng Ruan", "Zhiyu Fan", "Abhik Roychoudhury"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0332cfb0909325d94b5c927a0821873f32680a4f638119cd137a6251326203d", "submitted_at": "2024-04-08T11:55:09Z", "updated_at": "2024-07-25T16:54:41Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.05427v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.05427", "identifiers": {"arxiv": "2404.05427"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.05427"}], "name": "AutoCodeRover: Autonomous Program Improvement", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-25", "url": "https://arxiv.org/abs/2404.05427"}
{"aliases": [], "authors": ["Keen You", "Haotian Zhang", "Eldon Schoop", "Floris Weers", "Amanda Swearngin", "Jeffrey Nichols", "Yinfei Yang", "Zhe Gan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67e139ad83ab4804dfdc4c91f3fd344d7d3d279b97efb25bd11bab1532968e46", "submitted_at": "2024-04-08T17:55:44Z", "updated_at": "2024-04-08T17:55:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05719v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.05719", "identifiers": {"arxiv": "2404.05719"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.05719"}], "name": "Ferret-UI: Grounded Mobile UI Understanding with Multimodal LLMs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adapts multimodal modeling to mobile interfaces through detailed screen crops, region annotations, and UI reasoning tasks.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-04-08", "url": "https://arxiv.org/abs/2404.05719"}
{"aliases": [], "authors": ["Michael Lutz", "Arth Bohra", "Manvel Saroyan", "Artem Harutyunyan", "Giovanni Campagna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c18dbfedcccf6434b3a50a9aa62552372df8b772ccda6a1aa699a8f052476bc", "submitted_at": "2024-04-08T23:10:47Z", "updated_at": "2024-04-08T23:10:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05902v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2404.05902", "identifiers": {"arxiv": "2404.05902"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.05902"}], "name": "WILBUR: Adaptive In-Context Learning for Robust and Accurate Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "WILBUR learns to rank prior task demonstrations for a web agent’s prompt and synthesizes instructions from past runs. A generated task curriculum and a backtracking mechanism support adaptation and recovery, evaluated on WebVoyager.", "topics": ["web-navigation", "memory", "robustness", "data-generation"], "updated_at": "2024-04-08", "url": "https://arxiv.org/abs/2404.05902"}
{"aliases": [], "authors": ["Junpeng Liu", "Yifan Song", "Bill Yuchen Lin", "Wai Lam", "Graham Neubig", "Yuanzhi Li", "Xiang Yue"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7414bfa289baf86a8f2bb54f3f69f25a6a4cc8516fb0296a1899d21ad3020d4f", "submitted_at": "2024-04-09T02:29:39Z", "updated_at": "2024-04-09T02:29:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05955v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.05955", "identifiers": {"arxiv": "2404.05955"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.05955"}], "name": "VisualWebBench: How Far Have Multimodal LLMs Evolved in Web Page Understanding and Grounding?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separately evaluates web-page reading, understanding, and element grounding across tasks drawn from real websites.", "topics": ["evaluation", "screen-understanding", "grounding"], "updated_at": "2024-04-09", "url": "https://arxiv.org/abs/2404.05955"}
{"aliases": [], "authors": ["Shengding Hu", "Yuge Tu", "Xu Han", "Chaoqun He", "Ganqu Cui", "Xiang Long", "Zhi Zheng", "Yewei Fang", "Yuxiang Huang", "Weilin Zhao", "Xinrong Zhang", "Zheng Leng Thai", "Kaihuo Zhang", "Chongyi Wang", "Yuan Yao", "Chenyang Zhao", "Jie Zhou", "Jie Cai", "Zhongwu Zhai", "Ning Ding", "Chao Jia", "Guoyang Zeng", "Dahai Li", "Zhiyuan Liu", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bfd714e41bd7c6c53f1da1f7d9e04cd05abf6757934b8939928b630d6912fbd2", "submitted_at": "2024-04-09T15:36:50Z", "updated_at": "2024-06-03T08:54:38Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.06395v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2404.06395", "identifiers": {"arxiv": "2404.06395"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.06395"}], "name": "MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-03", "url": "https://arxiv.org/abs/2404.06395"}
{"aliases": [], "authors": ["Jiayi Pan", "Yichi Zhang", "Nicholas Tomlin", "Yifei Zhou", "Sergey Levine", "Alane Suhr"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "abc75ee55c7ac5992656dcec2106e286c670a58130a59b0eef7f5fd74d7496c5", "submitted_at": "2024-04-09T17:25:47Z", "updated_at": "2024-10-07T14:19:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.06474v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.06474", "identifiers": {"arxiv": "2404.06474"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.06474"}], "name": "Autonomous Evaluation and Refinement of Digital Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2024-04-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares automatic trajectory evaluators for web navigation and device control against oracle judgments. The resulting feedback is used for agent fine-tuning and inference-time guidance, including experiments on WebArena.", "topics": ["evaluation", "reward-models"], "updated_at": "2024-10-07", "url": "https://arxiv.org/abs/2404.06474"}
{"aliases": [], "authors": ["Tianbao Xie", "Danyang Zhang", "Jixuan Chen", "Xiaochuan Li", "Siheng Zhao", "Ruisheng Cao", "Toh Jing Hua", "Zhoujun Cheng", "Dongchan Shin", "Fangyu Lei", "Yitao Liu", "Yiheng Xu", "Shuyan Zhou", "Silvio Savarese", "Caiming Xiong", "Victor Zhong", "Tao Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a347bfcb4de03a60410392b52d522e5b7ad5dbabb474d60fc1ec097e74c4008d", "submitted_at": "2024-04-11T17:56:05Z", "updated_at": "2024-05-30T08:55:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.07972v2"}, "publication": {"authors": ["Tianbao Xie", "Danyang Zhang", "Jixuan Chen", "Xiaochuan Li", "Siheng Zhao", "Ruisheng Cao", "Toh J. Hua", "Zhoujun Cheng", "Dongchan Shin", "Fangyu Lei", "Yitao Liu", "Yiheng Xu", "Shuyan Zhou", "Silvio Savarese", "Caiming Xiong", "Victor Zhong", "Tao Yu"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-1650", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "52040-52094", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/5d413e48f84dc61244b6be550f1cd8f5-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["desktop", "linux", "windows", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "f62cbd2e5c92d250cc894a05e46f9052cfd38f0cdbf14902af6a96b96bcc9e8d", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.07972", "identifiers": {"arxiv": "2404.07972"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.07972"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/5d413e48f84dc61244b6be550f1cd8f5-Paper-Datasets_and_Benchmarks_Track.pdf"}], "name": "OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2024-04-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a real-computer evaluation environment with execution-based checks for tasks spanning desktop applications and operating-system state.", "topics": ["evaluation", "cross-app"], "updated_at": "2024-05-30", "url": "https://arxiv.org/abs/2404.07972"}
{"aliases": [], "authors": ["Nicolai Dorka", "Janusz Marecki", "Ammar Anwar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2f255b2c2870d9fbe5a543f2a0e09a30c3154fa0c13e9f46eedcefc11d2be43", "submitted_at": "2024-04-12T18:28:44Z", "updated_at": "2024-04-12T18:28:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.08755v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.08755", "identifiers": {"arxiv": "2404.08755"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.08755"}], "name": "Training a Vision Language Model as Smartphone Assistant", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains a smartphone assistant from sequences of screenshots paired with earlier actions. The model predicts UI-only interactions such as taps and swipes, with evaluation on Android in the Wild.", "topics": ["mobile-navigation", "interaction-history", "action-prediction"], "updated_at": "2024-04-12", "url": "https://arxiv.org/abs/2404.08755"}
{"aliases": [], "authors": ["Lei Ding", "Jeshwanth Bheemanpally", "Yi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b3a030a53a942eacfabafc64af67591fd0a85af66291c5d1b48baea14556e8aa", "submitted_at": "2024-04-13T00:20:09Z", "updated_at": "2024-07-09T01:14:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.08860v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.08860", "identifiers": {"arxiv": "2404.08860"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.08860"}], "name": "Enhancing Mobile \"How-to\" Queries with Automated Search Results Verification and Reranking", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-04-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Executes instructions from mobile how-to search results inside a controlled Android environment, then uses the observed success of those instructions to verify and rerank technical-support results.", "topics": ["mobile-navigation", "evaluation", "automation"], "updated_at": "2024-07-09", "url": "https://arxiv.org/abs/2404.08860"}
{"aliases": [], "authors": ["Chenhui Cui", "Tao Li", "Junjie Wang", "Chunyang Chen", "Dave Towey", "Rubing Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a844cdd06a32b8e89f362b7e924b36e27e72c1f1def81ed0b934f761ffcaba78", "submitted_at": "2024-04-13T09:56:50Z", "updated_at": "2026-06-25T14:05:55Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2404.08948v4"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.08948", "identifiers": {"arxiv": "2404.08948"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.08948"}], "name": "Large Language Models for Mobile GUI Text Input Generation: An Empirical Study", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies language-model generation of mobile text inputs, comparing UI representations, execution feedback, and human assistance for app testing.", "topics": ["automation", "evaluation"], "updated_at": "2026-06-25", "url": "https://arxiv.org/abs/2404.08948"}
{"aliases": [], "authors": ["Zhihao Cao", "Zidong Wang", "Siwen Xie", "Anji Liu", "Lifeng Fan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a065c845ac953c2ddfaa9a714b5ba72c18ede59145abb53eda12d587ad056703", "submitted_at": "2024-04-13T13:03:59Z", "updated_at": "2024-04-13T13:03:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.09001v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2404.09001", "identifiers": {"arxiv": "2404.09001"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.09001"}], "name": "Smart Help: Strategic Opponent Modeling for Proactive and Adaptive Robot Assistance in Households", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-13", "url": "https://arxiv.org/abs/2404.09001"}
{"aliases": [], "authors": ["Shulin Tian", "Ziniu Zhang", "Liangyu Chen", "Ziwei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "667b1f5053f3ae6ea712999932d5e9ee2f42e402cc72e713048d0790f5931976", "submitted_at": "2024-04-15T17:59:50Z", "updated_at": "2025-06-29T05:24:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.09992v2"}, "publication": {"authors": ["Shulin Tian", "Ziniu Zhang", "Liang-Yu Chen", "Ziwei Liu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.703", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.703/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "13682-13697", "pdf_url": "https://aclanthology.org/2025.findings-acl.703.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "MMInA: Benchmarking Multihop Multimodal Internet Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.703/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "5e725b77b94b484284adaab9ec70674c9fc13a7fce34b764e59561f9180b915f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.703/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.09992", "identifiers": {"arxiv": "2404.09992"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.09992"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.703/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.703.pdf"}], "name": "MMInA: Benchmarking Multihop Multimodal Internet Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates multimodal agents on compositional tasks across changing websites and studies replaying prior trajectories as memory support.", "topics": ["evaluation", "long-horizon", "memory"], "updated_at": "2025-06-29", "url": "https://arxiv.org/abs/2404.09992"}
{"aliases": [], "authors": ["SIMA Team", "Maria Abi Raad", "Arun Ahuja", "Catarina Barros", "Frederic Besse", "Andrew Bolt", "Adrian Bolton", "Bethanie Brownfield", "Gavin Buttimore", "Max Cant", "Sarah Chakera", "Stephanie C. Y. Chan", "Jeff Clune", "Adrian Collister", "Vikki Copeman", "Alex Cullum", "Ishita Dasgupta", "Dario de Cesare", "Julia Di Trapani", "Yani Donchev", "Emma Dunleavy", "Martin Engelcke", "Ryan Faulkner", "Frankie Garcia", "Charles Gbadamosi", "Zhitao Gong", "Lucy Gonzales", "Kshitij Gupta", "Karol Gregor", "Arne Olav Hallingstad", "Tim Harley", "Sam Haves", "Felix Hill", "Ed Hirst", "Drew A. Hudson", "Jony Hudson", "Steph Hughes-Fitt", "Danilo J. Rezende", "Mimi Jasarevic", "Laura Kampis", "Rosemary Ke", "Thomas Keck", "Junkyung Kim", "Oscar Knagg", "Kavya Kopparapu", "Rory Lawton", "Andrew Lampinen", "Shane Legg", "Alexander Lerchner", "Marjorie Limont", "Yulan Liu", "Maria Loks-Thompson", "Joseph Marino", "Kathryn Martin Cussons", "Loic Matthey", "Siobhan Mcloughlin", "Piermaria Mendolicchio", "Hamza Merzic", "Anna Mitenkova", "Alexandre Moufarek", "Valeria Oliveira", "Yanko Oliveira", "Hannah Openshaw", "Renke Pan", "Aneesh Pappu", "Alex Platonov", "Ollie Purkiss", "David Reichert", "John Reid", "Pierre Harvey Richemond", "Tyson Roberts", "Giles Ruscoe", "Jaume Sanchez Elias", "Tasha Sandars", "Daniel P. Sawyer", "Tim Scholtes", "Guy Simmons", "Daniel Slater", "Hubert Soyer", "Heiko Strathmann", "Peter Stys", "Allison C. Tam", "Denis Teplyashin", "Tayfun Terzi", "Davide Vercelli", "Bojan Vujatovic", "Marcus Wainwright", "Jane X. Wang", "Zhengdong Wang", "Daan Wierstra", "Duncan Williams", "Nathaniel Wong", "Sarah York", "Nick Young"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cbfdedf6a9130e395a06c97b46079dc7a158483cffebe8a8b2b6961f19ab5a46", "submitted_at": "2024-03-13T17:50:32Z", "updated_at": "2024-10-11T16:30:24Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.10179v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2404.10179", "identifiers": {"arxiv": "2404.10179"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.10179"}], "name": "Scaling Instructable Agents Across Many Simulated Worlds", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-11", "url": "https://arxiv.org/abs/2404.10179"}
{"aliases": [], "authors": ["Moghis Fereidouni", "Adib Mosharrof", "A. B. Siddique"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "127d037bb462ae4ed7bd72c46a62e2c4427611b30b67193454e40cc5c83cca3f", "submitted_at": "2024-04-16T20:15:32Z", "updated_at": "2025-01-26T19:07:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.10887v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.10887", "identifiers": {"arxiv": "2404.10887"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.10887"}], "name": "Grounded Language Agent for Product Search via Intelligent Web Interactions", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-04-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GLAINTEL, a Flan-T5-based product-search web agent studied under unsupervised learning, demonstration-based training, and unsupervised domain adaptation. It compares ways to combine demonstrations with reinforcement learning.", "topics": ["web-navigation", "reinforcement-learning", "imitation-learning"], "updated_at": "2025-01-26", "url": "https://arxiv.org/abs/2404.10887"}
{"aliases": [], "authors": ["Wei Chen", "Zhiyuan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9016b2bd0bfadbc17bbe6466a749a2d62f317fabe31e42e59d5bf9f7ed5b5d35", "submitted_at": "2024-04-17T15:07:06Z", "updated_at": "2026-07-20T02:21:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.11459v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.11459", "identifiers": {"arxiv": "2404.11459"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.11459"}], "name": "Octopus v3: Technical Report for On-device Sub-billion Multimodal AI Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-20", "url": "https://arxiv.org/abs/2404.11459"}
{"aliases": [], "authors": ["Chandeepa Dissanayake", "Lahiru Lowe", "Sachith Gunasekara", "Yasiru Ratnayake"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbd0451584b2c384252c1c9212eb6c6f3cfd35b622ba1976a3fe3419e08bc0e8", "submitted_at": "2024-04-18T13:57:18Z", "updated_at": "2024-04-18T13:57:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.12195v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2404.12195", "identifiers": {"arxiv": "2404.12195"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.12195"}], "name": "OpenBezoar: Small, Cost-Effective and Open Models Trained on Mixes of Instruction Data", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-18", "url": "https://arxiv.org/abs/2404.12195"}
{"aliases": [], "authors": ["Bertie Vidgen", "Adarsh Agrawal", "Ahmed M. Ahmed", "Victor Akinwande", "Namir Al-Nuaimi", "Najla Alfaraj", "Elie Alhajjar", "Lora Aroyo", "Trupti Bavalatti", "Max Bartolo", "Borhane Blili-Hamelin", "Kurt Bollacker", "Rishi Bomassani", "Marisa Ferrara Boston", "Siméon Campos", "Kal Chakra", "Canyu Chen", "Cody Coleman", "Zacharie Delpierre Coudert", "Leon Derczynski", "Debojyoti Dutta", "Ian Eisenberg", "James Ezick", "Heather Frase", "Brian Fuller", "Ram Gandikota", "Agasthya Gangavarapu", "Ananya Gangavarapu", "James Gealy", "Rajat Ghosh", "James Goel", "Usman Gohar", "Sujata Goswami", "Scott A. Hale", "Wiebke Hutiri", "Joseph Marvin Imperial", "Surgan Jandial", "Nick Judd", "Felix Juefei-Xu", "Foutse Khomh", "Bhavya Kailkhura", "Hannah Rose Kirk", "Kevin Klyman", "Chris Knotz", "Michael Kuchnik", "Shachi H. Kumar", "Srijan Kumar", "Chris Lengerich", "Bo Li", "Zeyi Liao", "Eileen Peters Long", "Victor Lu", "Sarah Luger", "Yifan Mai", "Priyanka Mary Mammen", "Kelvin Manyeki", "Sean McGregor", "Virendra Mehta", "Shafee Mohammed", "Emanuel Moss", "Lama Nachman", "Dinesh Jinenhally Naganna", "Amin Nikanjam", "Besmira Nushi", "Luis Oala", "Iftach Orr", "Alicia Parrish", "Cigdem Patlak", "William Pietri", "Forough Poursabzi-Sangdeh", "Eleonora Presani", "Fabrizio Puletti", "Paul Röttger", "Saurav Sahay", "Tim Santos", "Nino Scherrer", "Alice Schoenauer Sebag", "Patrick Schramowski", "Abolfazl Shahbazi", "Vin Sharma", "Xudong Shen", "Vamsi Sistla", "Leonard Tang", "Davide Testuggine", "Vithursan Thangarasa", "Elizabeth Anne Watkins", "Rebecca Weiss", "Chris Welty", "Tyler Wilbers", "Adina Williams", "Carole-Jean Wu", "Poonam Yadav", "Xianjun Yang", "Yi Zeng", "Wenhui Zhang", "Fedor Zhdanov", "Jiacheng Zhu", "Percy Liang", "Peter Mattson", "Joaquin Vanschoren"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bc2acfe4b74b1e6e412ccdfc0ad036c0dd6e9eb32316e1591f98699f1291fd51", "submitted_at": "2024-04-18T15:01:00Z", "updated_at": "2024-05-13T20:46:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.12241v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2404.12241", "identifiers": {"arxiv": "2404.12241"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.12241"}], "name": "Introducing v0.5 of the AI Safety Benchmark from MLCommons", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-13", "url": "https://arxiv.org/abs/2404.12241"}
{"aliases": [], "authors": ["Jason Wu", "Yi-Hao Peng", "Amanda Li", "Amanda Swearngin", "Jeffrey P. Bigham", "Jeffrey Nichols"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cfdf150f6f3d5e398216664af758070525812c6970b756add9233728951f1845", "submitted_at": "2024-04-18T20:43:08Z", "updated_at": "2024-04-18T20:43:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.12500v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2404.12500", "identifiers": {"arxiv": "2404.12500"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.12500"}], "name": "UIClip: A Data-driven Model for Assessing User Interface Design", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-18", "url": "https://arxiv.org/abs/2404.12500"}
{"aliases": [], "authors": ["Sibei Chen", "Yeye He", "Weiwei Cui", "Ju Fan", "Song Ge", "Haidong Zhang", "Dongmei Zhang", "Surajit Chaudhuri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c1e475bcf09106aab4e7757dd8858cf82a5fbbe14577798945cf1e91109a3f4b", "submitted_at": "2024-04-19T03:28:18Z", "updated_at": "2024-04-19T03:28:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.12608v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2404.12608", "identifiers": {"arxiv": "2404.12608"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.12608"}], "name": "Auto-Formula: Recommend Formulas in Spreadsheets using Contrastive Learning for Table Representations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-19", "url": "https://arxiv.org/abs/2404.12608"}
{"aliases": [], "authors": ["Wenhao Huang", "Zhouhong Gu", "Chenghao Peng", "Zhixu Li", "Jiaqing Liang", "Yanghua Xiao", "Liqian Wen", "Zulong Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9197968c51e668d3737496aaebadaa324e2b20b182aad541cabb491e491d3b41", "submitted_at": "2024-04-19T09:59:44Z", "updated_at": "2024-09-26T09:17:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.12753v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2404.12753", "identifiers": {"arxiv": "2404.12753"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.12753"}], "name": "AutoScraper: A Progressive Understanding Web Agent for Web Scraper Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-26", "url": "https://arxiv.org/abs/2404.12753"}
{"aliases": [], "authors": ["Konstantinos Vilouras", "Pedro Sanchez", "Alison Q. O'Neil", "Sotirios A. Tsaftaris"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f8519a7a8e3ba844292a202b4f3f9b73b25d10f798f48298992afac9ccdbfde", "submitted_at": "2024-04-19T14:43:48Z", "updated_at": "2025-01-30T16:31:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2404.12920v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2404.12920", "identifiers": {"arxiv": "2404.12920"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.12920"}], "name": "Zero-Shot Medical Phrase Grounding with Off-the-shelf Diffusion Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-30", "url": "https://arxiv.org/abs/2404.12920"}
{"aliases": [], "authors": ["Zhen Zeng", "William Watson", "Nicole Cho", "Saba Rahimi", "Shayleen Reynolds", "Tucker Balch", "Manuela Veloso"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c8fc04e8daa1bc15092e24f755abee6f963a34a568894ad3fb12ce3f9a092f6d", "submitted_at": "2024-03-17T00:36:37Z", "updated_at": "2024-03-17T00:36:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.13050v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2404.13050", "identifiers": {"arxiv": "2404.13050"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.13050"}], "name": "FlowMind: Automatic Workflow Generation with LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-03-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-03-17", "url": "https://arxiv.org/abs/2404.13050"}
{"aliases": [], "authors": ["Zeyu Zhang", "Xiaohe Bo", "Chen Ma", "Rui Li", "Xu Chen", "Quanyu Dai", "Jieming Zhu", "Zhenhua Dong", "Ji-Rong Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cfee9f61b20971e25eb26ba813b357e7f427361fbbe131479844c3b74f38ed5e", "submitted_at": "2024-04-21T01:49:46Z", "updated_at": "2024-04-21T01:49:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.13501v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2404.13501", "identifiers": {"arxiv": "2404.13501"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.13501"}], "name": "A Survey on the Memory Mechanism of Large Language Model based Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-21", "url": "https://arxiv.org/abs/2404.13501"}
{"aliases": [], "authors": ["Xiongzhi Zeng", "Huili Zhang", "Shizheng Zhang", "Pei Liu", "Kehuan Linghu", "Jiangyu Cui", "Xiaoxia Cai", "Jie Liu", "Zhenyu Li", "Jinlong Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "969d0fcf178e094491e29884995d0b180125125381a23cfe0c64d040827ea5e4", "submitted_at": "2024-04-22T09:52:50Z", "updated_at": "2024-04-22T09:52:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.14038v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2404.14038", "identifiers": {"arxiv": "2404.14038"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.14038"}], "name": "Accurate Chemical Reaction Modeling on Noisy Intermediate-Scale Quantum Computers Using a Noise-Resilient Wavefunction Ansatz", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-22", "url": "https://arxiv.org/abs/2404.14038"}
{"aliases": [], "authors": ["Ali Ebnenasir", "Kieran Young"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eb2dee25a109ac8a0a7d5b86751bbc5c5dbeceb058a3cb6029d15755a4abed99", "submitted_at": "2024-04-24T16:55:29Z", "updated_at": "2024-04-24T16:55:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.15980v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2404.15980", "identifiers": {"arxiv": "2404.15980"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.15980"}], "name": "Minimizing the Number of Teleportations in Distributed Quantum Computing Using Alloy", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-24", "url": "https://arxiv.org/abs/2404.15980"}
{"aliases": [], "authors": ["Kaining Ying", "Fanqing Meng", "Jin Wang", "Zhiqian Li", "Han Lin", "Yue Yang", "Hao Zhang", "Wenbo Zhang", "Yuqi Lin", "Shuo Liu", "Jiayi Lei", "Quanfeng Lu", "Runjian Chen", "Peng Xu", "Renrui Zhang", "Haozhe Zhang", "Peng Gao", "Yali Wang", "Yu Qiao", "Ping Luo", "Kaipeng Zhang", "Wenqi Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecc93699dd1d884d59407fddc042bd5ce1801331fb30c4a7057564cd9cc66fbd", "submitted_at": "2024-04-24T17:37:05Z", "updated_at": "2024-04-24T17:37:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.16006v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.16006", "identifiers": {"arxiv": "2404.16006"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.16006"}], "name": "MMT-Bench: A Comprehensive Multimodal Benchmark for Evaluating Large Vision-Language Models Towards Multitask AGI", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-04-24", "url": "https://arxiv.org/abs/2404.16006"}
{"aliases": [], "authors": ["Rajat Chawla", "Adarsh Jha", "Muskaan Kumar", "Mukunda NS", "Ishaan Bhola"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e29d14a41d59942c88938b8e83217484a1cb619847c5dff0ec1f516095ee5654", "submitted_at": "2024-04-09T11:59:41Z", "updated_at": "2024-10-27T05:54:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.16048v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.16048", "identifiers": {"arxiv": "2404.16048"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.16048"}], "name": "GUIDE: Graphical User Interface Data for Execution", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces annotated website task trajectories with screenshots, prior actions, reasoning, and grounded next actions for process automation.", "topics": ["data-generation", "grounding"], "updated_at": "2024-10-27", "url": "https://arxiv.org/abs/2404.16048"}
{"aliases": [], "authors": ["Li Zhang", "Shihe Wang", "Xianqing Jia", "Zhihan Zheng", "Yunhe Yan", "Longxi Gao", "Yuanchun Li", "Mengwei Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "28c3ad086106ec4199dcc9da2b8aa3b20eac9347d9354b9dfb373e50ea4c3f50", "submitted_at": "2024-04-12T15:39:09Z", "updated_at": "2024-08-02T13:49:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.16054v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.16054", "identifiers": {"arxiv": "2404.16054"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.16054"}], "name": "LlamaTouch: A Faithful and Scalable Testbed for Mobile UI Task Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "LlamaTouch executes mobile tasks on devices and evaluates whether an agent visits manually annotated essential application or system states. Screenshot and hierarchy annotations, combined with exact and fuzzy state matching, support evaluation of 496 tasks across multiple agents.", "topics": ["evaluation", "mobile-navigation", "state-tracking"], "updated_at": "2024-08-02", "url": "https://arxiv.org/abs/2404.16054"}
{"aliases": [], "authors": ["Juyong Lee", "Taywon Min", "Minyong An", "Dongyoon Hahm", "Haeone Lee", "Changyeon Kim", "Kimin Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45c13064bbfbcf153566b86c8969f4de5b5e87dc133342c595b3b58bde3bb615", "submitted_at": "2024-04-25T14:56:32Z", "updated_at": "2025-07-21T02:55:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.16660v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.16660", "identifiers": {"arxiv": "2404.16660"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.16660"}], "name": "Benchmarking Mobile Device Control Agents across Diverse Configurations", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-04-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces B-MoCA, an interactive mobile-control benchmark that varies device settings, layouts, and languages to test generalization.", "topics": ["evaluation", "robustness"], "updated_at": "2025-07-21", "url": "https://arxiv.org/abs/2404.16660"}
{"aliases": [], "authors": ["Zirui Song", "Yaohang Li", "Meng Fang", "Yanda Li", "Zhenhao Chen", "Zecheng Shi", "Yuan Huang", "Xiuying Chen", "Ling Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e531ec2ef9e5d90586b8f3052e16d41da95114b7382047d0a02fd671a0db355", "submitted_at": "2024-04-28T05:33:15Z", "updated_at": "2025-03-23T13:04:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.18074v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.18074", "identifiers": {"arxiv": "2404.18074"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.18074"}], "name": "MMAC-Copilot: Multi-modal Agent Collaboration Operating Copilot", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2024-04-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MMAC-Copilot, which coordinates agents with different expertise to operate PC applications. VIBench evaluates visual interaction with applications that lack usable APIs, spanning office, recreation and game tasks.", "topics": ["automation", "evaluation", "planning"], "updated_at": "2025-03-23", "url": "https://arxiv.org/abs/2404.18074"}
{"aliases": [], "authors": ["Wei Chen", "Zhiyuan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f1ba53df7744bc7d6da5d6edb05e33e821343bc74b919576cd1c6db4d17111a", "submitted_at": "2024-04-30T06:55:45Z", "updated_at": "2026-07-20T02:16:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.19296v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2404.19296", "identifiers": {"arxiv": "2404.19296"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2404.19296"}], "name": "Octopus v4: Graph of language models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-04-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-20", "url": "https://arxiv.org/abs/2404.19296"}
{"aliases": [], "authors": ["Lucas-Andreï Thil", "Mirela Popa", "Gerasimos Spanakis"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3498a7bc289c6537ab69780a56a9fd0a2bb697fb74c077949824c4fc018762c6", "submitted_at": "2024-05-01T13:51:45Z", "updated_at": "2024-05-01T13:51:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.00516v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2405.00516", "identifiers": {"arxiv": "2405.00516"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.00516"}], "name": "Navigating WebAI: Training Agents to Complete Web Tasks with Large Language Models and Reinforcement Learning", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-05-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines supervised and reinforcement learning for MiniWoB web tasks and investigates whether agents understand HTML structure or memorize target elements.", "topics": ["web-navigation", "reinforcement-learning", "evaluation"], "updated_at": "2024-05-01", "url": "https://arxiv.org/abs/2405.00516"}
{"aliases": [], "authors": ["Michael Wornow", "Avanika Narayan", "Krista Opsahl-Ong", "Quinn McIntyre", "Nigam H. Shah", "Christopher Re"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e19c81e6b6000f3e609fbc8c77aab7be6e4e231d4ee467d7924d07e336df2017", "submitted_at": "2024-05-03T23:25:15Z", "updated_at": "2024-05-03T23:25:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.03710v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2405.03710", "identifiers": {"arxiv": "2405.03710"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.03710"}], "name": "Automating the Enterprise with Foundation Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-03", "url": "https://arxiv.org/abs/2405.03710"}
{"aliases": [], "authors": ["Razan Baltaji", "Babak Hemmatian", "Lav R. Varshney"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "751e0c6747e87889ac7a7a56bbee05cc1bc058c8abd0a92473a0c2a490d6a028", "submitted_at": "2024-05-06T21:20:35Z", "updated_at": "2024-08-14T18:01:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2405.03862v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2405.03862", "identifiers": {"arxiv": "2405.03862"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.03862"}], "name": "Persona Inconstancy in Multi-Agent LLM Collaboration: Conformity, Confabulation, and Impersonation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-14", "url": "https://arxiv.org/abs/2405.03862"}
{"aliases": [], "authors": ["Kihoon Son", "Jinhyeon Kwon", "DaEun Choi", "Tae Soo Kim", "Young-Ho Kim", "Sangdoo Yun", "Juho Kim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a358d417d8bef89a01e799d88564e3ede4678ebda4950e87c96cf1e1954b85e1", "submitted_at": "2024-05-07T17:10:31Z", "updated_at": "2024-05-08T05:44:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.04497v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2405.04497", "identifiers": {"arxiv": "2405.04497"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.04497"}], "name": "Unveiling Disparities in Web Task Handling Between Human and Web Agent", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-05-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares human think-aloud web-task behavior with web-agent planning, action, and reflection. The analysis highlights differences in updating plans, handling ambiguity, and investigating failures.", "topics": ["evaluation", "human-in-the-loop", "planning"], "updated_at": "2024-05-08", "url": "https://arxiv.org/abs/2405.04497"}
{"aliases": [], "authors": ["Matthew Renze", "Erhan Guven"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "334ac570d7cfc00616ac9bd4d78273d7a27cd169b7d2d1afd732ff8571f5d498", "submitted_at": "2024-05-05T18:56:46Z", "updated_at": "2024-10-16T23:19:46Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2405.06682v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2405.06682", "identifiers": {"arxiv": "2405.06682"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.06682"}], "name": "Self-Reflection in LLM Agents: Effects on Problem-Solving Performance", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-16", "url": "https://arxiv.org/abs/2405.06682"}
{"aliases": [], "authors": ["Sidong Feng", "Suyu Ma", "Han Wang", "David Kong", "Chunyang Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d200a997596ea919d5210b5ca5760241f9dfa3ad581e090a541871cd6e117522", "submitted_at": "2024-05-11T20:32:16Z", "updated_at": "2024-05-11T20:32:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.07090v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2405.07090", "identifiers": {"arxiv": "2405.07090"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.07090"}], "name": "MUD: Towards a Large-Scale and Noise-Filtered UI Dataset for Modern Style UI Modeling", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2024-05-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "MUD combines LLM-guided Android app exploration, noise filtering and human annotation to collect modern UI screenshots. The study evaluates the resulting data for element detection and UI retrieval.", "topics": ["data-generation", "screen-understanding"], "updated_at": "2024-05-11", "url": "https://arxiv.org/abs/2405.07090"}
{"aliases": [], "authors": ["Dionysia Danai Brilli", "Evangelos Georgaras", "Stefania Tsilivaki", "Nikos Melanitis", "Konstantina Nikita"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "44ee4ab198e8800daecfb1dd9c66915fe039df956fbe6f6d1a048213ccb395a2", "submitted_at": "2024-05-13T10:09:37Z", "updated_at": "2024-08-09T19:54:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.07606v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2405.07606", "identifiers": {"arxiv": "2405.07606"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.07606"}], "name": "AIris: An AI-powered Wearable Assistive Device for the Visually Impaired", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-09", "url": "https://arxiv.org/abs/2405.07606"}
{"aliases": [], "authors": ["Sivaprasad Omanakuttan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ed5c7083362c3a9b5d659f2560030f4be488aa3a83f4c77aa5832e2358bf6ec6", "submitted_at": "2024-05-13T16:19:31Z", "updated_at": "2024-05-13T16:19:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.07885v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2405.07885", "identifiers": {"arxiv": "2405.07885"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.07885"}], "name": "Quantum Computation Using Large Spin Qudits", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-05-13", "url": "https://arxiv.org/abs/2405.07885"}
{"aliases": [], "authors": ["Wanting Xu", "Yang Liu", "Langping He", "Xucheng Huang", "Ling Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d52cc1d83bf929fdaa3a42e641e2d75ac499a93503482382918c24ac11f9cde9", "submitted_at": "2024-05-15T09:47:59Z", "updated_at": "2024-06-20T07:31:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2405.09215v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2405.09215", "identifiers": {"arxiv": "2405.09215"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.09215"}], "name": "Xmodel-VLM: A Simple Baseline for Multimodal Vision Language Model", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-20", "url": "https://arxiv.org/abs/2405.09215"}
{"aliases": [], "authors": ["Yuexiang Zhai", "Hao Bai", "Zipeng Lin", "Jiayi Pan", "Shengbang Tong", "Yifei Zhou", "Alane Suhr", "Saining Xie", "Yann LeCun", "Yi Ma", "Sergey Levine"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "450d05efac6a9d09107b80d494b01da8c0473dd41738248edf9bfff69b755990", "submitted_at": "2024-05-16T17:50:19Z", "updated_at": "2024-10-07T19:13:47Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2405.10292v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2405.10292", "identifiers": {"arxiv": "2405.10292"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.10292"}], "name": "Fine-Tuning Large Vision-Language Models as Decision-Making Agents via Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-07", "url": "https://arxiv.org/abs/2405.10292"}
{"aliases": [], "authors": ["William E Bishop", "Alice Li", "Christopher Rawles", "Oriana Riva"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81f62326dddc5c512fb05472eb1f387064c0a5aad77453e9d2e85415bb627fa9", "submitted_at": "2024-05-17T23:27:33Z", "updated_at": "2024-05-17T23:27:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.11120v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2405.11120", "identifiers": {"arxiv": "2405.11120"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.11120"}], "name": "Latent State Estimation Helps UI Agents to Reason", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-05-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies explicit textual estimates of latent UI state, including whether commanded actions occurred and how far a task has progressed. The experiments compare agents that reason with these estimates against agents that omit them across public and internal task benchmarks.", "topics": ["state-tracking", "planning", "evaluation"], "updated_at": "2024-05-17", "url": "https://arxiv.org/abs/2405.11120"}
{"aliases": [], "authors": ["Arushi Jain", "Shubham Paliwal", "Monika Sharma", "Lovekesh Vig", "Gautam Shroff"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ca8d7e7dc3a537573ac3990d532589cdca12bf441ee40521cc46044fd10c5263", "submitted_at": "2024-05-21T14:49:12Z", "updated_at": "2024-05-21T14:49:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.12842v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2405.12842", "identifiers": {"arxiv": "2405.12842"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.12842"}], "name": "SmartFlow: Robotic Process Automation using LLMs", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2024-05-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents SmartFlow, which converts visible GUI elements into text for language-model action planning and executes the plan through a scripting engine. A collection of enterprise applications evaluates adaptation to different layouts and tasks.", "topics": ["automation", "workflow", "screen-understanding"], "updated_at": "2024-05-21", "url": "https://arxiv.org/abs/2405.12842"}
{"aliases": [], "authors": ["Christopher Rawles", "Sarah Clinckemaillie", "Yifan Chang", "Jonathan Waltz", "Gabrielle Lau", "Marybeth Fair", "Alice Li", "William Bishop", "Wei Li", "Folawiyo Campbell-Ajala", "Daniel Toyama", "Robert Berry", "Divya Tyamagundlu", "Timothy Lillicrap", "Oriana Riva"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a33b8b0a03c0dc7ac3f699c34193a526d20791bfedfdde2a2fb89604959e08c2", "submitted_at": "2024-05-23T13:48:54Z", "updated_at": "2025-04-06T20:37:50Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2405.14573v5"}, "publication": {"authors": ["Chris Rawles", "Sarah Clinckemaillie", "Yifan Chang", "Jonathan Waltz", "Gabrielle Lau", "Marybeth Fair", "Alice Li", "William Bishop", "Wei Li", "Folawiyo Campbell-Ajala", "Daniel Toyama", "Robert Berry", "Divya Tyamagundlu", "Timothy Lillicrap", "Oriana Riva"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "406-441", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/01a83bc2f2732a58e6aa731e659e7101-Paper-Conference.pdf", "publisher": null, "series": null, "title": "AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:00Z", "role": "primary", "sha256": "e525e5eeedcb08e61765eb0d4034c43381bd3ecc5cc4bf9cc8788cf50e7a2bca", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2405.14573", "identifiers": {"arxiv": "2405.14573"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.14573"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/01a83bc2f2732a58e6aa731e659e7101-Paper-Conference.pdf"}], "name": "AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2024-05-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces parameterized Android tasks with initialization, success checks, and cleanup for reproducible interactive evaluation.", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2025-04-06", "url": "https://arxiv.org/abs/2405.14573"}
{"aliases": [], "authors": ["Peiyuan Feng", "Yichen He", "Guanhua Huang", "Yuan Lin", "Hanchong Zhang", "Yuchen Zhang", "Hang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4cd4c5736e291dd7799681e33bbea0cfec630997be3b36ba078eb151cba7eded", "submitted_at": "2024-05-23T16:17:44Z", "updated_at": "2024-11-05T09:42:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.14751v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2405.14751", "identifiers": {"arxiv": "2405.14751"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.14751"}], "name": "AGILE: A Novel Reinforcement Learning Framework of LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-05", "url": "https://arxiv.org/abs/2405.14751"}
{"aliases": [], "authors": ["Weiyu Li", "Jiarui Liu", "Hongyu Yan", "Rui Chen", "Yixun Liang", "Xuelin Chen", "Ping Tan", "Xiaoxiao Long"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2388989c2371fa9695ae878cf22063ecc20a22397c9d937b17d01b49556820c1", "submitted_at": "2024-05-23T18:30:12Z", "updated_at": "2025-05-30T10:22:03Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2405.14979v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2405.14979", "identifiers": {"arxiv": "2405.14979"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.14979"}], "name": "CraftsMan3D: High-fidelity Mesh Generation with 3D Native Generation and Interactive Geometry Refiner", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-30", "url": "https://arxiv.org/abs/2405.14979"}
{"aliases": [], "authors": ["Abdur Rahman", "Rajat Chawla", "Muskaan Kumar", "Arkajit Datta", "Adarsh Jha", "Mukunda NS", "Ishaan Bhola"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7cb6087a48596442e1d191bf1313e9459ad62c02c3104ee481cb72cb00e56b1", "submitted_at": "2024-05-24T08:21:45Z", "updated_at": "2024-07-21T07:34:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.15341v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:27Z", "role": "primary", "sha256": "92875dc9298b7952cc6e33ce1165b806421a6e24f3cd1e9bfd951fdd928659ad", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2405.15341v2"}], "id": "arxiv:2405.15341", "identifiers": {"arxiv": "2405.15341"}, "kind": "paper", "last_verified_at": "2026-10-17T09:13:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.15341"}], "name": "V-Zen: Efficient GUI Understanding and Precise Grounding With A Novel Multimodal LLM", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced. The related GUIDE dataset paper remains a separate record; this annotation describes the V-Zen model contribution.", "published_at": "2024-05-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes V-Zen, a multimodal GUI model with dual-resolution image encoders for grounding and next-action prediction, fine-tuned using GUIDE interface elements and task sequences.", "topics": ["grounding", "action-prediction", "screen-understanding"], "updated_at": "2024-07-21", "url": "https://arxiv.org/abs/2405.15341"}
{"aliases": [], "authors": ["John Yang", "Carlos E. Jimenez", "Alexander Wettig", "Kilian Lieret", "Shunyu Yao", "Karthik Narasimhan", "Ofir Press"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea972f3764419aa209152b1bb8ba9b7d2688da1cb009cbdd4aa62664c263911a", "submitted_at": "2024-05-06T17:41:33Z", "updated_at": "2024-11-11T20:01:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2405.15793v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2405.15793", "identifiers": {"arxiv": "2405.15793"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.15793"}], "name": "SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-11", "url": "https://arxiv.org/abs/2405.15793"}
{"aliases": [], "authors": ["Minghao Chen", "Yihang Li", "Yanting Yang", "Shiyu Yu", "Binbin Lin", "Xiaofei He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d2c250456b40ddd3c6dbe09aafe3377e13144b6b334cd629e3c4c4a62f6fc58", "submitted_at": "2024-05-25T14:11:44Z", "updated_at": "2024-11-10T12:54:35Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2405.16247v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2405.16247", "identifiers": {"arxiv": "2405.16247"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2405.16247"}], "name": "AutoManual: Constructing Instruction Manuals by LLM Agents via Interactive Environmental Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-05-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-10", "url": "https://arxiv.org/abs/2405.16247"}
{"aliases": [], "authors": ["Ajay Patel", "Markus Hofmarcher", "Claudiu Leoveanu-Condrei", "Marius-Constantin Dinu", "Chris Callison-Burch", "Sepp Hochreiter"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77b292052a2ab4e8db61606ffe1da8bd5e63b3ce8c84b2e0e4b7a2fdc2f926ca", "submitted_at": "2024-05-30T17:52:36Z", "updated_at": "2024-10-01T21:28:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.20309v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2405.20309", "identifiers": {"arxiv": "2405.20309"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.20309"}], "name": "Large Language Models Can Self-Improve At Web Agent Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-05-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Examines self-improvement by fine-tuning web agents on three mixtures of model-generated training data. WebArena experiments and trajectory-level evaluation measures study task performance, robustness, and behavior beyond aggregate success rates.", "topics": ["data-generation", "web-navigation", "evaluation"], "updated_at": "2024-10-01", "url": "https://arxiv.org/abs/2405.20309"}
{"aliases": [], "authors": ["Junyang Wang", "Haiyang Xu", "Haitao Jia", "Xi Zhang", "Ming Yan", "Weizhou Shen", "Ji Zhang", "Fei Huang", "Jitao Sang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a1338f3543a04bc2f9734e5fcabbfd5ebda6ac263461c2305f7c39383e938573", "submitted_at": "2024-06-03T05:50:00Z", "updated_at": "2024-06-03T05:50:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.01014v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2406.01014", "identifiers": {"arxiv": "2406.01014"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.01014"}], "name": "Mobile-Agent-v2: Mobile Device Operation Assistant with Effective Navigation via Multi-Agent Collaboration", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates mobile planning, decision-making, and reflection, with a memory unit for task progress and relevant content.", "topics": ["planning", "memory", "mobile-navigation"], "updated_at": "2024-06-03", "url": "https://arxiv.org/abs/2406.01014"}
{"aliases": [], "authors": ["Richard Zhipeng Wang", "James S. Cummins", "Marvin Syed", "Nikita Stroev", "George Pastras", "Jason Sakellariou", "Symeon Tsintzos", "Alexis Askitopoulos", "Daniele Veraldi", "Marcello Calvanese Strinati", "Silvia Gentilini", "Davide Pierangeli", "Claudio Conti", "Natalia G. Berloff"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b8347ba74a375970bcf070c5a390ab3fceddfc255234024089b1ad5fb7fb6960", "submitted_at": "2024-06-03T15:03:31Z", "updated_at": "2024-06-03T15:03:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.01400v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2406.01400", "identifiers": {"arxiv": "2406.01400"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.01400"}], "name": "Efficient Computation Using Spatial-Photonic Ising Machines: Utilizing Low-Rank and Circulant Matrix Constraints", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-03", "url": "https://arxiv.org/abs/2406.01400"}
{"aliases": [], "authors": ["Eric Li", "Jim Waldo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9a728edf0b834adb72abfc50e13d5b2b52f532f444b3e8af34effd3d86beb1b", "submitted_at": "2024-06-01T00:32:26Z", "updated_at": "2024-06-01T00:32:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.01623v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2406.01623", "identifiers": {"arxiv": "2406.01623"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.01623"}], "name": "WebSuite: Systematically Evaluating Why Web Agents Fail", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebSuite organizes browser operations into an action taxonomy and tests both isolated controls and complete workflows. Its diagnostic tasks attribute failures to specific actions, enabling comparison of text-based and multimodal web agents.", "topics": ["evaluation", "observability", "web-navigation"], "updated_at": "2024-06-01", "url": "https://arxiv.org/abs/2406.01623"}
{"aliases": [], "authors": ["Yichuan Wang", "Yang Liu", "Yu Yan", "Qun Wang", "Xucheng Huang", "Ling Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cf9a85f509aec7b65f882bd2ba8591463d6483f061124071f1aadd77132df06d", "submitted_at": "2024-06-05T02:12:06Z", "updated_at": "2024-11-19T08:38:55Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2406.02856v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2406.02856", "identifiers": {"arxiv": "2406.02856"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.02856"}], "name": "Xmodel-LM Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-19", "url": "https://arxiv.org/abs/2406.02856"}
{"aliases": [], "authors": ["Robert-Jeron Reifert", "Hayssam Dahrouj", "Alaa Alameer Ahmad", "Haris Gacanin", "Aydin Sezgin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b0048b1a0718832e6cb40e06a9d150297c5e34f9d2929566bb3e011f91eff3a5", "submitted_at": "2024-06-05T18:00:09Z", "updated_at": "2024-06-05T18:00:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.03548v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2406.03548", "identifiers": {"arxiv": "2406.03548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.03548"}], "name": "Robust Communication and Computation using Deep Learning via Joint Uncertainty Injection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-05", "url": "https://arxiv.org/abs/2406.03548"}
{"aliases": [], "authors": ["Wei Li", "William Bishop", "Alice Li", "Chris Rawles", "Folawiyo Campbell-Ajala", "Divya Tyamagundlu", "Oriana Riva"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6eb3209d8e2a431987fe19f01cd0818f9fb490e684d7369e7d3002dc388ff271", "submitted_at": "2024-06-06T01:49:29Z", "updated_at": "2024-11-13T16:42:22Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2406.03679v6"}, "publication": {"authors": ["Wei Li", "William Bishop", "Alice Li", "Chris Rawles", "Folawiyo Campbell-Ajala", "Divya Tyamagundlu", "Oriana Riva"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-2925", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "92130-92154", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/a79f3ef3b445fd4659f44648f7ea8ffd-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "On the Effects of Data Scale on UI Control Agents", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:03Z", "role": "primary", "sha256": "0f499dc380fcd787579aab846e3d1e39b720b672fd9916b4a016688218ca48dd", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2406.03679", "identifiers": {"arxiv": "2406.03679"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.03679"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/a79f3ef3b445fd4659f44648f7ea8ffd-Paper-Datasets_and_Benchmarks_Track.pdf"}], "name": "On the Effects of Data Scale on UI Control Agents", "notes": "Annotation reviewed against the primary abstract. Original dataset scale and experimental results are not independently reproduced.", "published_at": "2024-06-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies how Android task performance changes with demonstration scale and releases AndroidControl with high- and low-level instructions.", "topics": ["data-generation", "mobile-navigation", "evaluation"], "updated_at": "2024-11-13", "url": "https://arxiv.org/abs/2406.03679"}
{"aliases": [], "authors": ["Yubo Dong", "Xukun Zhu", "Zhengzhe Pan", "Linchao Zhu", "Yi Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "433d4543f0f4a3a23dfff71f44c3233d940169380a63d531cfec052041cd2720", "submitted_at": "2024-06-09T10:21:47Z", "updated_at": "2024-06-09T10:21:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.05720v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2406.05720", "identifiers": {"arxiv": "2406.05720"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.05720"}], "name": "VillagerAgent: A Graph-Based Multi-Agent Framework for Coordinating Complex Task Dependencies in Minecraft", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-09", "url": "https://arxiv.org/abs/2406.05720"}
{"aliases": [], "authors": ["Junhee Cho", "Jihoon Kim", "Daseul Bae", "Jinho Choo", "Youngjune Gwon", "Yeong-Dae Kwon"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f85b2cf7650f41ba9ee6e4ae83cd9041e077a44adcb685e702d644bd5779963", "submitted_at": "2024-06-11T05:21:20Z", "updated_at": "2024-12-26T11:02:28Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.06947v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2406.06947", "identifiers": {"arxiv": "2406.06947"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.06947"}], "name": "CAAP: Context-Aware Action Planning Prompting to Solve Computer Tasks with Front-End UI Only", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "CAAP converts screenshots into textual observations and plans keyboard-and-mouse actions using a prompt that examines task context from multiple perspectives. The agent is evaluated on MiniWoB++ and WebShop using only front-end visual observations.", "topics": ["planning", "screen-understanding", "web-navigation"], "updated_at": "2024-12-26", "url": "https://arxiv.org/abs/2406.06947"}
{"aliases": [], "authors": ["Benhao Huang", "Yingzhuo Yu", "Jin Huang", "Xingjian Zhang", "Jiaqi Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63c50612018569a90c5ce81532d87189089be234809ec92e39d4fc6ee40a8166", "submitted_at": "2024-06-11T14:02:23Z", "updated_at": "2025-05-26T18:23:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.07275v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2406.07275", "identifiers": {"arxiv": "2406.07275"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.07275"}], "name": "DCA-Bench: A Benchmark for Dataset Curation Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-26", "url": "https://arxiv.org/abs/2406.07275"}
{"aliases": [], "authors": ["Andrea Burns", "Kate Saenko", "Bryan A. Plummer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47d0ca0467a104f9b8be037ec73131f12b66a22b0a89fe975bec7680073c8f39", "submitted_at": "2024-06-12T02:43:19Z", "updated_at": "2024-08-08T02:36:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.07822v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2406.07822", "identifiers": {"arxiv": "2406.07822"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.07822"}], "name": "Tell Me What's Next: Textual Foresight for Generic UI Representations", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Textual Foresight learns UI representations by describing the future screen expected after an action on the current interface. The OpenApp dataset supports pretraining and downstream evaluation of representations that connect individual elements with whole-screen changes.", "topics": ["screen-understanding", "world-models", "data-generation"], "updated_at": "2024-08-08", "url": "https://arxiv.org/abs/2406.07822"}
{"aliases": [], "authors": ["Luyuan Wang", "Yongyu Deng", "Yiwei Zha", "Guodong Mao", "Qinmin Wang", "Tianchen Min", "Wei Chen", "Shoufa Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "637843b1b8634e608bed71fc0f3df205ae218d1b0b5550ce480046c67bb14b0f", "submitted_at": "2024-06-12T13:14:50Z", "updated_at": "2024-06-12T13:14:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.08184v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2406.08184", "identifiers": {"arxiv": "2406.08184"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.08184"}], "name": "MobileAgentBench: An Efficient and User-Friendly Benchmark for Mobile LLM Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a mobile-agent benchmark with tasks across open-source applications, organized by difficulty to reduce manual evaluation effort.", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2024-06-12", "url": "https://arxiv.org/abs/2406.08184"}
{"aliases": [], "authors": ["Shengcheng Yu", "Chunrong Fang", "Mingzhe Du", "Zimin Ding", "Zhenyu Chen", "Zhendong Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0061e9004afe11bf188164ea37de1acd15e1b95ae2fcad22044264a89e4636b3", "submitted_at": "2024-06-12T15:48:39Z", "updated_at": "2024-06-12T15:48:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.08340v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2406.08340", "identifiers": {"arxiv": "2406.08340"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.08340"}], "name": "Practical, Automated Scenario-based Mobile App Testing", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "ScenTest builds scenario-specific event knowledge graphs from crowdsourced test reports containing GUI and action information. It uses the current app state and testing context to generate scenario tests across applications.", "topics": ["automation", "cross-app", "screen-understanding", "evaluation"], "updated_at": "2024-06-12", "url": "https://arxiv.org/abs/2406.08340"}
{"aliases": [], "authors": ["Quanfeng Lu", "Wenqi Shao", "Zitao Liu", "Lingxiao Du", "Fanqing Meng", "Boxuan Li", "Botong Chen", "Siyuan Huang", "Kaipeng Zhang", "Ping Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b45039151353cc0fede71e2e9ba109ec53d0c1f495037a82d8fb742c9e97bbb", "submitted_at": "2024-06-12T17:44:26Z", "updated_at": "2025-08-01T03:37:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.08451v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2406.08451", "identifiers": {"arxiv": "2406.08451"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.08451"}], "name": "GUIOdyssey: A Comprehensive Dataset for Cross-App GUI Navigation on Mobile Devices", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces annotated cross-application mobile trajectories and studies history-aware navigation over long task sequences.", "topics": ["cross-app", "data-generation"], "updated_at": "2025-08-01", "url": "https://arxiv.org/abs/2406.08451"}
{"aliases": [], "authors": ["Zhen Xiang", "Linzhi Zheng", "Yanjie Li", "Junyuan Hong", "Qinbin Li", "Han Xie", "Jiawei Zhang", "Zidi Xiong", "Chulin Xie", "Carl Yang", "Dawn Song", "Bo Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50a8ea492d3f1c2ff2bd00ebfb18e19077bacc96b44c751d985bfe0f7bef0247", "submitted_at": "2024-06-13T14:49:26Z", "updated_at": "2025-05-29T03:09:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.09187v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2406.09187", "identifiers": {"arxiv": "2406.09187"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.09187"}], "name": "GuardAgent: Safeguard LLM Agents by a Guard Agent via Knowledge-Enabled Reasoning", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The web-agent evidence is Mind2Web-SC; the healthcare benchmark does not independently establish GUI control.", "published_at": "2024-06-13", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces GuardAgent, which translates safety requests into executable guardrail code using plans and retrieved demonstrations. Mind2Web-SC supplies a web-agent safety-policy evaluation alongside a separate healthcare access-control benchmark.", "topics": ["safety-security", "evaluation", "tool-integration"], "updated_at": "2025-05-29", "url": "https://arxiv.org/abs/2406.09187"}
{"aliases": [], "authors": ["Zhe Yuan", "Jianqi Shi", "Yanhong Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "287e9ad320d5ec7261daecf373554d4d2cdd1f2f4e43f45f0947a26da3af92af", "submitted_at": "2024-06-14T10:47:52Z", "updated_at": "2024-08-06T08:27:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.09913v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}], "id": "arxiv:2406.09913", "identifiers": {"arxiv": "2406.09913"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.09913"}], "name": "OpenECAD: An Efficient Visual Language Model for Editable 3D-CAD Design", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-06", "url": "https://arxiv.org/abs/2406.09913"}
{"aliases": [], "authors": ["Kevin Qinghong Lin", "Linjie Li", "Difei Gao", "Qinchen WU", "Mingyi Yan", "Zhengyuan Yang", "Lijuan Wang", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "66b5fb1e839490df80e9750fdc049f7fe3d2d4686c4090d78ea01bc542fe55b0", "submitted_at": "2024-06-14T17:59:08Z", "updated_at": "2024-06-14T17:59:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.10227v1"}, "publication": {"authors": ["Kevin Q. Lin", "Linjie Li", "Difei Gao", "Qinchen Wu", "Mingyi Yan", "Zhengyuan Yang", "Lijuan Wang", "Mike Z. Shou"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-2214", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "69329-69360", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/804e757b7d7043c26701c3a313032101-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "VideoGUI: A Benchmark for GUI Automation from Instructional Videos", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:03Z", "role": "primary", "sha256": "5731b34872eb84e83338aa851d7f3922e304162869753bf6100ffdc1183c96cf", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2406.10227", "identifiers": {"arxiv": "2406.10227"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.10227"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/804e757b7d7043c26701c3a313032101-Paper-Datasets_and_Benchmarks_Track.pdf"}], "name": "VideoGUI: A Benchmark for GUI Automation from Instructional Videos", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates GUI automation from instructional videos at separate levels of high-level planning, action narration, and primitive execution.", "topics": ["evaluation", "video", "planning"], "updated_at": "2024-06-14", "url": "https://arxiv.org/abs/2406.10227"}
{"aliases": [], "authors": ["Rithesh Murthy", "Liangwei Yang", "Juntao Tan", "Tulika Manoj Awalgaonkar", "Yilun Zhou", "Shelby Heinecke", "Sachin Desai", "Jason Wu", "Ran Xu", "Sarah Tan", "Jianguo Zhang", "Zhiwei Liu", "Shirley Kokane", "Zuxin Liu", "Ming Zhu", "Huan Wang", "Caiming Xiong", "Silvio Savarese"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f6e943eab6221971dee78326d54fc64a95a6d1c29a42ef3224ae8949b95efee3", "submitted_at": "2024-06-12T22:58:12Z", "updated_at": "2024-06-12T22:58:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.10290v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.10290", "identifiers": {"arxiv": "2406.10290"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.10290"}], "name": "MobileAIBench: Benchmarking LLMs and LMMs for On-Device Use Cases", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-12", "url": "https://arxiv.org/abs/2406.10290"}
{"aliases": [], "authors": ["Avinash Maurya", "Robert Underwood", "M. Mustafa Rafique", "Franck Cappello", "Bogdan Nicolae"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "19413767b757b10fd7a79cb7eae88191c5922c1cec15d17f3d237879553c904f", "submitted_at": "2024-06-15T18:30:40Z", "updated_at": "2024-06-15T18:30:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.10707v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.10707", "identifiers": {"arxiv": "2406.10707"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.10707"}], "name": "DataStates-LLM: Lazy Asynchronous Checkpointing for Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-15", "url": "https://arxiv.org/abs/2406.10707"}
{"aliases": [], "authors": ["Dongping Chen", "Yue Huang", "Siyuan Wu", "Jingyu Tang", "Liuyi Chen", "Yilin Bai", "Zhigang He", "Chenlong Wang", "Huichi Zhou", "Yiqiang Li", "Tianshuo Zhou", "Yue Yu", "Chujie Gao", "Qihui Zhang", "Yi Gui", "Zhen Li", "Yao Wan", "Pan Zhou", "Jianfeng Gao", "Lichao Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9997c3ab43f0e3e69ad4dcbe15257ea4392d36fa3c82f90c6b7730667a067f99", "submitted_at": "2024-06-16T06:56:53Z", "updated_at": "2025-03-24T11:46:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.10819v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2406.10819", "identifiers": {"arxiv": "2406.10819"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.10819"}], "name": "GUI-World: A Video Benchmark and Dataset for Multimodal GUI-oriented Understanding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces video-based GUI questions and annotations for studying temporal interface understanding across varied interaction settings.", "topics": ["evaluation", "video", "screen-understanding"], "updated_at": "2025-03-24", "url": "https://arxiv.org/abs/2406.10819"}
{"aliases": [], "authors": ["Wentong Chen", "Junbo Cui", "Jinyi Hu", "Yujia Qin", "Junjie Fang", "Yue Zhao", "Chongyi Wang", "Jun Liu", "Guirong Chen", "Yupeng Huo", "Yuan Yao", "Yankai Lin", "Zhiyuan Liu", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40e41704dbe59f1fb9b1e743a3c13394105de112213f666c2420fb235d666ef3", "submitted_at": "2024-06-17T08:30:55Z", "updated_at": "2025-05-30T07:30:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.11317v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2406.11317", "identifiers": {"arxiv": "2406.11317"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.11317"}], "name": "GUICourse: From General Vision Language Models to Versatile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUIEnv, GUIAct, and GUIChat datasets to develop OCR, grounding, component knowledge, and interaction abilities in visual-language models.", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-05-30", "url": "https://arxiv.org/abs/2406.11317"}
{"aliases": [], "authors": ["Zhao Zhuo", "Rongzhen Li", "Kai Liu", "Huhai Zou", "KaiMao Li", "Jie Yu", "Tianhao Sun", "Qingbo Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5260b51d98495bbd128e32af8fd3c6327cf53ccad8aacc251c7e442762e4cf8a", "submitted_at": "2024-06-17T08:59:32Z", "updated_at": "2024-08-09T01:45:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.11342v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.11342", "identifiers": {"arxiv": "2406.11342"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.11342"}], "name": "KAOS: Large Model Multi-Agent Operating System", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-09", "url": "https://arxiv.org/abs/2406.11342"}
{"aliases": [], "authors": ["Jieyu Zhang", "Weikai Huang", "Zixian Ma", "Oscar Michel", "Dong He", "Tanmay Gupta", "Wei-Chiu Ma", "Ali Farhadi", "Aniruddha Kembhavi", "Ranjay Krishna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4b32dd32bc566e458346242abe354579f14edb16e56ea0e438949284d634bec5", "submitted_at": "2024-06-17T17:32:42Z", "updated_at": "2025-01-27T06:25:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.11775v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.11775", "identifiers": {"arxiv": "2406.11775"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.11775"}], "name": "Task Me Anything", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-27", "url": "https://arxiv.org/abs/2406.11775"}
{"aliases": [], "authors": ["Hao Bai", "Yifei Zhou", "Mert Cemri", "Jiayi Pan", "Alane Suhr", "Sergey Levine", "Aviral Kumar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c5d4e75d6ba88574f6f033bb2c1cf44cf04149e94e017cd8a1a227e7a1c8238", "submitted_at": "2024-06-14T17:49:55Z", "updated_at": "2024-06-14T17:49:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.11896v1"}, "publication": {"authors": ["Hao Bai", "Yifei Zhou", "Mert Cemri", "Jiayi Pan", "Alane Suhr", "Sergey Levine", "Aviral Kumar"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-0397", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "12461-12495", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/1704ddd0bb89f159dfe609b32c889995-Paper-Conference.pdf", "publisher": null, "series": null, "title": "DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:03Z", "role": "primary", "sha256": "0fd68a0f088591eb695e205286823c7be4059fba8a14e62030ca1b5b11707e21", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "arxiv:2406.11896", "identifiers": {"arxiv": "2406.11896"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.11896"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/1704ddd0bb89f159dfe609b32c889995-Paper-Conference.pdf"}], "name": "DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines offline initialization and online reinforcement learning in Android environments, with automated evaluation and curriculum-based sampling.", "topics": ["reinforcement-learning", "mobile-navigation"], "updated_at": "2024-06-14", "url": "https://arxiv.org/abs/2406.11896"}
{"aliases": [], "authors": ["Shunyu Yao", "Noah Shinn", "Pedram Razavi", "Karthik Narasimhan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "966ffa41610a1be69eb5feeb1daa23d83fd30aa7f3884091ce0063dad7f1982e", "submitted_at": "2024-06-17T19:33:08Z", "updated_at": "2024-06-17T19:33:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.12045v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.12045", "identifiers": {"arxiv": "2406.12045"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.12045"}], "name": "$τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-17", "url": "https://arxiv.org/abs/2406.12045"}
{"aliases": [], "authors": ["Xubing Ye", "Yukang Gan", "Xiaoke Huang", "Yixiao Ge", "Yansong Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "625b9b5209958a498cf729b70f7bb528fe5ab62d349b579bff6022bf32873390", "submitted_at": "2024-06-18T05:05:12Z", "updated_at": "2025-03-03T09:05:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.12275v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.12275", "identifiers": {"arxiv": "2406.12275"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.12275"}], "name": "VoCo-LLaMA: Towards Vision Compression with Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-03", "url": "https://arxiv.org/abs/2406.12275"}
{"aliases": [], "authors": ["Yichen Pan", "Dehan Kong", "Sida Zhou", "Cheng Cui", "Yifei Leng", "Bing Jiang", "Hangyu Liu", "Yanyi Shang", "Shuyan Zhou", "Tongshuang Wu", "Zhengyang Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5d0fbea7a4813a1da85c9ecb237c2115cc183b4c9e6fd0cbfd63764b0d418e3", "submitted_at": "2024-06-18T07:58:33Z", "updated_at": "2024-07-16T06:19:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.12373v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2406.12373", "identifiers": {"arxiv": "2406.12373"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.12373"}], "name": "WebCanvas: Benchmarking Web Agents in Online Environments", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebCanvas and Mind2Web-Live, evaluating online task progress through meaningful intermediate states on changing websites.", "topics": ["evaluation", "web-navigation"], "updated_at": "2024-07-16", "url": "https://arxiv.org/abs/2406.12373"}
{"aliases": [], "authors": ["Xuan Zhang", "Yang Deng", "Zifeng Ren", "See-Kiong Ng", "Tat-Seng Chua"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2aec6a074d7abb8d7c9c71f78ff481a7af23d90f586baa3e9633fcf9cb62022b", "submitted_at": "2024-06-18T14:07:28Z", "updated_at": "2024-10-02T02:02:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.12639v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2406.12639", "identifiers": {"arxiv": "2406.12639"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.12639"}], "name": "Ask-before-Plan: Proactive Language Agents for Real-World Planning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-02", "url": "https://arxiv.org/abs/2406.12639"}
{"aliases": [], "authors": ["Chen Henry Wu", "Rishi Shah", "Jing Yu Koh", "Ruslan Salakhutdinov", "Daniel Fried", "Aditi Raghunathan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cc1c1ac93f8e328d33accf7cdaa1d65487c4f7abec15234b9e596a501947550", "submitted_at": "2024-06-18T17:32:48Z", "updated_at": "2025-02-04T20:02:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.12814v3"}, "publication": {"authors": ["Chen Wu", "Rishi Shah", "Jing Yu Koh", "Russ Salakhutdinov", "Daniel Fried", "Aditi Raghunathan"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "28362-28383", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/460a1d8eac34125dad453b28d6d64446-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Dissecting Adversarial Robustness of Multimodal LM Agents", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:01Z", "role": "primary", "sha256": "fa23fda1b80dac019741737fc6f611fccf9234a4d73ffbb8e24e54d8caf4dac3", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2406.12814", "identifiers": {"arxiv": "2406.12814"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.12814"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/460a1d8eac34125dad453b28d6d64446-Paper-Conference.pdf"}], "name": "Dissecting Adversarial Robustness of Multimodal LM Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2024-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces adversarial tasks on VisualWebArena and an evaluation framework that tracks how manipulated information passes between agent components. The analysis examines vulnerabilities introduced by reflection and tree-search evaluators as well as direct action generation.", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2025-02-04", "url": "https://arxiv.org/abs/2406.12814"}
{"aliases": [], "authors": ["Suriya Ganesh Ayyamperumal", "Limin Ge"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08c3ff630ed8408727b39e97903e0bd3b9646f353db6f10875b3d590623967a1", "submitted_at": "2024-06-16T22:04:10Z", "updated_at": "2024-06-16T22:04:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.12934v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.12934", "identifiers": {"arxiv": "2406.12934"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.12934"}], "name": "Current state of LLM Risks and AI Guardrails", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-16", "url": "https://arxiv.org/abs/2406.12934"}
{"aliases": [], "authors": ["Michael Wornow", "Avanika Narayan", "Ben Viggiano", "Ishan S. Khare", "Tathagat Verma", "Tibor Thompson", "Miguel Angel Fuentes Hernandez", "Sudharsan Sundar", "Chloe Trujillo", "Krrish Chawla", "Rongfei Lu", "Justin Shen", "Divya Nagaraj", "Joshua Martinez", "Vardhan Agrawal", "Althea Hudson", "Nigam H. Shah", "Christopher Re"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7e7d00030ea8c77b9ed6f950d62f2bd417d3870f24b2f4589d8068c9484e328", "submitted_at": "2024-06-19T06:50:15Z", "updated_at": "2024-10-11T00:06:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.13264v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:30Z", "role": "primary", "sha256": "95c08b3a28f77a248d6f990345ef4b2b16cbf196492920bfa1ce0ec7e398345e", "supports": ["name", "kind", "summary", "topics", "notes"], "url": "https://arxiv.org/abs/2406.13264v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:16:02Z", "role": "primary", "sha256": "07de7eac76877364a2d9e9d31f6c65f5f2cace2f97f7821862a1441f83d9f95c", "supports": ["summary", "environments", "scope", "notes"], "url": "https://arxiv.org/html/2406.13264v2"}], "id": "arxiv:2406.13264", "identifiers": {"arxiv": "2406.13264"}, "kind": "paper", "last_verified_at": "2026-10-17T09:15:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.13264"}], "name": "WONDERBREAD: A Benchmark for Evaluating Multimodal Foundation Models on Business Process Management Tasks", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced. The demonstrations were rehearsed, so their completion rate is not a zero-shot human performance estimate.", "published_at": "2024-06-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WONDERBREAD, a benchmark pairing human WebArena workflow demonstrations with screen recordings, action logs, and written procedures. Six evaluation tasks cover documentation, knowledge transfer, and process improvement beyond end-to-end execution.", "topics": ["evaluation", "workflow", "video"], "updated_at": "2024-10-11", "url": "https://arxiv.org/abs/2406.13264"}
{"aliases": [], "authors": ["Edoardo Debenedetti", "Jie Zhang", "Mislav Balunović", "Luca Beurer-Kellner", "Marc Fischer", "Florian Tramèr"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "598c0ce1fcc190026d03953faf3c3f675f1e759d8de3ffc52c3bc1dad8191d22", "submitted_at": "2024-06-19T08:55:56Z", "updated_at": "2024-11-24T22:04:23Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.13352v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.13352", "identifiers": {"arxiv": "2406.13352"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.13352"}], "name": "AgentDojo: A Dynamic Environment to Evaluate Prompt Injection Attacks and Defenses for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-24", "url": "https://arxiv.org/abs/2406.13352"}
{"aliases": [], "authors": ["Qinchen Wu", "Difei Gao", "Kevin Qinghong Lin", "Zhuoyu Wu", "Xiangwu Guo", "Peiran Li", "Weichen Zhang", "Hengxu Wang", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "369c32dba4bab0cfc05b41db16c73d5672605b7f81caabb2dbd72f581b3290ec", "submitted_at": "2024-06-19T17:22:11Z", "updated_at": "2024-06-19T17:22:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.13719v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2406.13719", "identifiers": {"arxiv": "2406.13719"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.13719"}], "name": "GUI Action Narrator: Where and When Did That Action Take Place?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Act2Cap and a cursor-guided method for describing GUI actions from video, with temporal and spatial localization challenges.", "topics": ["video", "screen-understanding"], "updated_at": "2024-06-19", "url": "https://arxiv.org/abs/2406.13719"}
{"aliases": [], "authors": ["Ziyang Meng", "Yu Dai", "Zezheng Gong", "Shaoxiong Guo", "Minglong Tang", "Tongquan Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea674b3aec3aee51c6f7fd4e99ad2b88394e7df43a4f3a02654d805e9a60343a", "submitted_at": "2024-06-20T07:24:43Z", "updated_at": "2024-11-04T09:31:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.14056v3"}, "publication": {"authors": ["Meng Ziyang", "Yu Dai", "Zezheng Gong", "Shaoxiong Guo", "Minglong Tang", "Tongquan Wei"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.68", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.68/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "1261-1279", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.68.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "VGA: Vision GUI Assistant - Minimizing Hallucinations through Image-Centric Fine-Tuning", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.68/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "7202817c9f68a7c61afa104997fffe2a1a1112977561309ff5811f722b9411e8", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-emnlp.68/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2406.14056", "identifiers": {"arxiv": "2406.14056"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.14056"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-emnlp.68/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-emnlp.68.pdf"}], "name": "VGA: Vision GUI Assistant -- Minimizing Hallucinations through Image-Centric Fine-Tuning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses image-dependent question-answer data and staged fine-tuning to improve GUI comprehension and reduce unsupported visual answers.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-11-04", "url": "https://arxiv.org/abs/2406.14056"}
{"aliases": [], "authors": ["Ke Wang", "Tianyu Xia", "Zhangxuan Gu", "Yi Zhao", "Shuheng Shen", "Changhua Meng", "Weiqiang Wang", "Ke Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4d5a46d2141ca59fcf221bbe618601cb7cce29c1b067ae60712e53e907095bd", "submitted_at": "2024-06-20T12:22:05Z", "updated_at": "2024-07-01T08:37:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.14250v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2406.14250", "identifiers": {"arxiv": "2406.14250"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.14250"}], "name": "E-ANT: A Large-Scale Dataset for Efficient Automatic GUI NavigaTion", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Chinese-language GUI navigation traces and annotated screenshots collected from human interactions with mini-applications.", "topics": ["mobile-navigation", "evaluation"], "updated_at": "2024-07-01", "url": "https://arxiv.org/abs/2406.14250"}
{"aliases": [], "authors": ["Omri Berkovitch", "Sapir Caduri", "Noam Kahlon", "Anatoly Efros", "Avi Caciularu", "Ido Dagan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0bc798b69e59d1ec0ca8ecb0fab7a7ee6594f4d656ca4212b19be7e3085c0c5", "submitted_at": "2024-06-20T13:46:10Z", "updated_at": "2025-03-03T15:47:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.14314v3"}}, "environments": ["android", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2406.14314", "identifiers": {"arxiv": "2406.14314"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.14314"}], "name": "Identifying User Goals from UI Trajectories", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Defines goal identification from observed UI trajectories and an evaluation method for comparing intent descriptions within the same interface context. Experiments reuse Android and web automation data to compare human and model inference of users’ detailed goals.", "topics": ["evaluation", "interaction-history", "screen-understanding"], "updated_at": "2025-03-03", "url": "https://arxiv.org/abs/2406.14314"}
{"aliases": [], "authors": ["Gabriel Sarch", "Lawrence Jang", "Michael J. Tarr", "William W. Cohen", "Kenneth Marino", "Katerina Fragkiadaki"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "698e2c444a587a109bbc7b3d64b812a3d0cf2ef7148ff02e3e6fe13f19f193fd", "submitted_at": "2024-06-20T17:45:02Z", "updated_at": "2025-09-18T02:44:34Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2406.14596v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2406.14596", "identifiers": {"arxiv": "2406.14596"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.14596"}], "name": "VLM Agents Generate Their Own Memories: Distilling Experience into Embodied Programs of Thought", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-18", "url": "https://arxiv.org/abs/2406.14596"}
{"aliases": [], "authors": ["Ruixuan Xiao", "Wentao Ma", "Ke Wang", "Yuchuan Wu", "Junbo Zhao", "Haobo Wang", "Fei Huang", "Yongbin Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b2663746d70aa80cba068e183211af4da75447a9c4cb96a1e6273217dc3b76a8", "submitted_at": "2024-06-21T06:13:00Z", "updated_at": "2024-06-21T06:13:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.14884v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.14884", "identifiers": {"arxiv": "2406.14884"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.14884"}], "name": "FlowBench: Revisiting and Benchmarking Workflow-Guided Planning for LLM-based Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-21", "url": "https://arxiv.org/abs/2406.14884"}
{"aliases": [], "authors": ["Zeyao Ma", "Bohan Zhang", "Jing Zhang", "Jifan Yu", "Xiaokang Zhang", "Xiaohan Zhang", "Sijia Luo", "Xi Wang", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0b967db7f54e65df2497abe2681d2bb63c2571d2ef8624f6f5af76766f43f487", "submitted_at": "2024-06-21T09:06:45Z", "updated_at": "2024-10-17T07:23:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.14991v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.14991", "identifiers": {"arxiv": "2406.14991"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.14991"}], "name": "SpreadsheetBench: Towards Challenging Real World Spreadsheet Manipulation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-17", "url": "https://arxiv.org/abs/2406.14991"}
{"aliases": [], "authors": ["Wei Chen", "Zhiyuan Li", "Zhen Guo", "Yikang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cd08cd743581099a5c2d346f2b227d5ace220ffa7df1281e5d6b5cd7a06edc1c", "submitted_at": "2024-06-26T05:40:10Z", "updated_at": "2026-07-20T02:13:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.18082v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2406.18082", "identifiers": {"arxiv": "2406.18082"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.18082"}], "name": "Octo-planner: On-device Language Model for Planner-Action Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-20", "url": "https://arxiv.org/abs/2406.18082"}
{"aliases": [], "authors": ["Xinyu Lian", "Sam Ade Jacobs", "Lev Kurilenko", "Masahiro Tanaka", "Stas Bekman", "Olatunji Ruwase", "Minjia Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6418efa56cb187a65af5779748ca29b508e66545cd517fad1f0c9b815aa5cb5c", "submitted_at": "2024-06-27T01:28:30Z", "updated_at": "2025-07-04T06:16:30Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.18820v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.18820", "identifiers": {"arxiv": "2406.18820"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.18820"}], "name": "Universal Checkpointing: A Flexible and Efficient Distributed Checkpointing System for Large-Scale DNN Training with Reconfigurable Parallelis", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-04", "url": "https://arxiv.org/abs/2406.18820"}
{"aliases": [], "authors": ["Yue Fan", "Lei Ding", "Ching-Chen Kuo", "Shan Jiang", "Yang Zhao", "Xinze Guan", "Jie Yang", "Yi Zhang", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0c86d68b5c3486b099f64792d5d3290d79e25b7c82c6eb2402869b893bb22fea", "submitted_at": "2024-06-27T15:34:16Z", "updated_at": "2024-10-25T18:16:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.19263v2"}, "publication": {"authors": ["Yue Fan", "Lei Ding", "Ching-Chen Kuo", "Shan Jiang", "Yang Zhao", "Xinze Guan", "Jie Yang", "Yi Zhang", "Xin Eric Wang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.emnlp-main.533", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.emnlp-main.533/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "9503-9522", "pdf_url": "https://aclanthology.org/2024.emnlp-main.533.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Read Anywhere Pointed: Layout-aware GUI Screen Reading with Tree-of-Lens Grounding", "type": "conference-paper", "url": "https://aclanthology.org/2024.emnlp-main.533/", "venue": "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "1d105799f5db853e801090e8ed62273cd72dd3cb9b4060f901747aace8d2406d", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.emnlp-main.533/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2406.19263", "identifiers": {"arxiv": "2406.19263"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.19263"}, {"rel": "publication", "url": "https://aclanthology.org/2024.emnlp-main.533/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.emnlp-main.533.pdf"}], "name": "Read Anywhere Pointed: Layout-aware GUI Screen Reading with Tree-of-Lens Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-06-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Screen Point-and-Read and a hierarchical layout approach that interprets screen regions indicated by a user-selected point.", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-10-25", "url": "https://arxiv.org/abs/2406.19263"}
{"aliases": [], "authors": ["Noel Crawford", "Edward B. Duffy", "Iman Evazzade", "Torsten Foehr", "Gregory Robbins", "Debbrata Kumar Saha", "Jiya Varma", "Marcin Ziolkowski"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d789a22c68e775f31fe130b2dd8fb312dfe8088e24ffe02ddd8f4e7594a6b18d", "submitted_at": "2024-06-28T16:39:20Z", "updated_at": "2024-07-02T11:45:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.20041v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.20041", "identifiers": {"arxiv": "2406.20041"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.20041"}], "name": "BMW Agents -- A Framework For Task Automation Through Multi-Agent Collaboration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-02", "url": "https://arxiv.org/abs/2406.20041"}
{"aliases": [], "authors": ["Jieneng Chen", "Luoxin Ye", "Ju He", "Zhao-Yang Wang", "Daniel Khashabi", "Alan Yuille"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "585ef95fc14cc24a60446e81a4906cc193a8aed653334355472a2efbdede2607", "submitted_at": "2024-06-28T17:57:14Z", "updated_at": "2024-11-17T17:05:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.20092v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2406.20092", "identifiers": {"arxiv": "2406.20092"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2406.20092"}], "name": "Efficient Large Multi-modal Models via Visual Context Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-17", "url": "https://arxiv.org/abs/2406.20092"}
{"aliases": [], "authors": ["Kai Siong Yow", "Chunbo Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7eb10229911369b613f4af832e34e06d3ede8424ef536ad473f1da7623c3ec3a", "submitted_at": "2024-06-18T15:04:16Z", "updated_at": "2024-06-18T15:04:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.00068v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2407.00068", "identifiers": {"arxiv": "2407.00068"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.00068"}], "name": "D&A: Resource Optimisation in Personalised PageRank Computations Using Multi-Core Machines", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-18", "url": "https://arxiv.org/abs/2407.00068"}
{"aliases": [], "authors": ["Franz G. Fuchs", "Alexander J. Stasik", "Stanley Miao", "Ola Tangen Kulseng", "Ruben Pariente Bassa"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f6e1e382909d57ab488fdb7a323c34a9df0f04d43ee379efc723b938932b0113", "submitted_at": "2024-06-29T13:59:51Z", "updated_at": "2024-06-29T13:59:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.00445v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2407.00445", "identifiers": {"arxiv": "2407.00445"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.00445"}], "name": "Quantum reservoir computing using the stabilizer formalism for encoding classical data", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-06-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-06-29", "url": "https://arxiv.org/abs/2407.00445"}
{"aliases": [], "authors": ["Shihan Deng", "Weikai Xu", "Hongda Sun", "Wei Liu", "Tao Tan", "Jianfeng Liu", "Ang Li", "Jian Luan", "Bin Wang", "Rui Yan", "Shuo Shang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "90ce9900644c21004e0ba9e749e10d94238d2447d0e84bd7205ff55dcbd243f4", "submitted_at": "2024-07-01T06:10:01Z", "updated_at": "2024-07-01T06:10:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.00993v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.00993", "identifiers": {"arxiv": "2407.00993"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.00993"}], "name": "Mobile-Bench: An Evaluation Benchmark for LLM-based Mobile Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates mobile agents using combined UI and API actions, tasks with varying app scope, and checkpoints for intermediate progress.", "topics": ["evaluation", "cross-app", "tool-integration"], "updated_at": "2024-07-01", "url": "https://arxiv.org/abs/2407.00993"}
{"aliases": [], "authors": ["Jing Yu Koh", "Stephen McAleer", "Daniel Fried", "Ruslan Salakhutdinov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ec7aaf2a34e38f72da652bb2d8350c951229c154e2bc368a927a79287b36c439", "submitted_at": "2024-07-01T17:07:55Z", "updated_at": "2026-02-08T15:06:40Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2407.01476v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.01476", "identifiers": {"arxiv": "2407.01476"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.01476"}], "name": "Tree Search for Language Model Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2024-07-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Applies best-first search to agent actions within interactive web environments, allowing exploration and multistep planning at inference time. WebArena and VisualWebArena experiments examine search behavior and test-time computation.", "topics": ["planning", "web-navigation", "evaluation"], "updated_at": "2026-02-08", "url": "https://arxiv.org/abs/2407.01476"}
{"aliases": [], "authors": ["Chunqiu Steven Xia", "Yinlin Deng", "Soren Dunn", "Lingming Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6941fbccf6fac666381089814d39373d3a6c905bb8cb5cf16573bef8ed39fc40", "submitted_at": "2024-07-01T17:24:45Z", "updated_at": "2024-10-29T17:29:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.01489v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.01489", "identifiers": {"arxiv": "2407.01489"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.01489"}], "name": "Agentless: Demystifying LLM-based Software Engineering Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-29", "url": "https://arxiv.org/abs/2407.01489"}
{"aliases": [], "authors": ["Sayash Kapoor", "Benedikt Stroebl", "Zachary S. Siegel", "Nitya Nadgir", "Arvind Narayanan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c67699cfc21fc5ce4bafdaa57dd5b1191b89095bd86a4dbfcc066fef6833e26", "submitted_at": "2024-07-01T17:48:14Z", "updated_at": "2024-07-01T17:48:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.01502v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.01502", "identifiers": {"arxiv": "2407.01502"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.01502"}], "name": "AI Agents That Matter", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-01", "url": "https://arxiv.org/abs/2407.01502"}
{"aliases": [], "authors": ["Tianqi Xu", "Linyao Chen", "Dai-Jie Wu", "Yanjun Chen", "Zecheng Zhang", "Xiang Yao", "Zhiqiang Xie", "Yongchao Chen", "Shilong Liu", "Bochen Qian", "Anjie Yang", "Zhaoxuan Jin", "Jianbo Deng", "Philip Torr", "Bernard Ghanem", "Guohao Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "12e5464ea0fc62bd566f19d301b887826b7ff317cfc27c86fa0cda8d3de271d9", "submitted_at": "2024-07-01T17:55:04Z", "updated_at": "2025-07-20T13:42:07Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2407.01511v4"}, "publication": {"authors": ["Tianqi Xu", "Linyao Chen", "Dai-Jie Wu", "Yanjun Chen", "Zecheng Zhang", "Xiang Yao", "Zhiqiang Xie (谢志强)", "Yongchao Chen", "Shilong Liu", "Bochen Qian", "Anjie Yang", "Zhaoxuan Jin", "Jianbo Deng", "Philip Torr", "Bernard Ghanem", "Guohao Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.1113", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.1113/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "21607-21647", "pdf_url": "https://aclanthology.org/2025.findings-acl.1113.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "CRAB: Cross-environment Agent Benchmark for Multimodal Language Model Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.1113/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "a8cf178f2faa088fd94b77d03785b8c6e2908db34b411287984066b4e5d1ef3f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.1113/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.01511", "identifiers": {"arxiv": "2407.01511"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.01511"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.1113/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.1113.pdf"}], "name": "CRAB: Cross-environment Agent Benchmark for Multimodal Language Model Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CRAB, a framework for cross-environment agent tasks with graph-based evaluation and a benchmark spanning desktop and phone interactions.", "topics": ["evaluation", "cross-app"], "updated_at": "2025-07-20", "url": "https://arxiv.org/abs/2407.01511"}
{"aliases": [], "authors": ["El Hassane Ettifouri", "Jessica López Espejel", "Laura Minkova", "Tassnim Dardouri", "Walid Dahhane"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "01d3b3a013f3b10249b8197a8e18aa695bdc20febaa34f07128791f36afdcbac", "submitted_at": "2024-05-05T19:10:19Z", "updated_at": "2025-07-18T11:35:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.01558v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.01558", "identifiers": {"arxiv": "2407.01558"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.01558"}], "name": "Visual Grounding Methods for Efficient Interaction with Desktop Graphical User Interfaces", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-05-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares modular OCR-based and end-to-end methods for locating desktop GUI targets from natural-language instructions.", "topics": ["grounding", "evaluation"], "updated_at": "2025-07-18", "url": "https://arxiv.org/abs/2407.01558"}
{"aliases": [], "authors": ["E. Manuylovich", "A. E. Bednyakova", "D. A. Ivoilov", "I. S. Terekhov", "S. K. Turitsyn"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9fbad2b1110de23b08d29e8bf2417dff8b199170e97af36e37b17142abf63d06", "submitted_at": "2024-07-02T10:05:58Z", "updated_at": "2024-07-02T10:05:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.02117v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2407.02117", "identifiers": {"arxiv": "2407.02117"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.02117"}], "name": "SOA-based reservoir computing using up-sampling", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-02", "url": "https://arxiv.org/abs/2407.02117"}
{"aliases": [], "authors": ["Zhe Liu", "Cheng Li", "Chunyang Chen", "Junjie Wang", "Mengzhuo Chen", "Boyu Wu", "Yawen Wang", "Jun Hu", "Qing Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73b706ac6089a7d19836268437f3fea58d10b9d170ac596ac0d5a12f0d3a9668", "submitted_at": "2024-07-03T11:58:09Z", "updated_at": "2024-12-04T08:00:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.03037v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.03037", "identifiers": {"arxiv": "2407.03037"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.03037"}], "name": "Seeing is Believing: Vision-driven Non-crash Functional Bug Detection for Mobile Apps", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trident combines exploration, monitoring, and detection agents to identify functional bugs that do not crash a mobile app. It reasons over screenshot sequences and interface transitions to construct test oracles, evaluated on 590 non-crash bugs and additional live apps.", "topics": ["automation", "evaluation", "screen-understanding", "state-tracking"], "updated_at": "2024-12-04", "url": "https://arxiv.org/abs/2407.03037"}
{"aliases": [], "authors": ["Arindam Mitra", "Luciano Del Corro", "Guoqing Zheng", "Shweti Mahajan", "Dany Rouhana", "Andres Codas", "Yadong Lu", "Wei-ge Chen", "Olga Vrousgos", "Corby Rosset", "Fillipe Silva", "Hamed Khanpour", "Yash Lara", "Ahmed Awadallah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a299e4173c3451798682f6e6bcd3a7edbcd443123b156603522c66fec3686bef", "submitted_at": "2024-07-03T21:01:12Z", "updated_at": "2024-07-03T21:01:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.03502v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.03502", "identifiers": {"arxiv": "2407.03502"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.03502"}], "name": "AgentInstruct: Toward Generative Teaching with Agentic Flows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-03", "url": "https://arxiv.org/abs/2407.03502"}
{"aliases": [], "authors": ["Jiayi Zhang", "Chuang Zhao", "Yihan Zhao", "Zhaoyang Yu", "Ming He", "Jianping Fan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa74755c53e20b3a9ff4b72e03a557f455ecceaefc96a50587aefd78ac3c8176", "submitted_at": "2024-07-04T13:12:19Z", "updated_at": "2024-07-04T13:12:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.03913v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.03913", "identifiers": {"arxiv": "2407.03913"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.03913"}], "name": "MobileExperts: A Dynamic Tool-Enabled Agent Team in Mobile Devices", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-07-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MobileExperts, which assembles mobile-agent teams, lets agents explore and formulate tools, and coordinates them through two levels of planning. A benchmark organizes tasks by hierarchical intelligence levels.", "topics": ["mobile-navigation", "planning", "tool-integration"], "updated_at": "2024-07-04", "url": "https://arxiv.org/abs/2407.03913"}
{"aliases": [], "authors": ["Songqin Nong", "Jiali Zhu", "Rui Wu", "Jiongchao Jin", "Shuo Shan", "Xiutian Huang", "Wenhao Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "59708ce4edeba4e913ae890f458b51907dd2f3b0b10bb18c86469b36375c84e4", "submitted_at": "2024-07-05T08:37:10Z", "updated_at": "2024-12-06T07:43:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.04346v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2407.04346", "identifiers": {"arxiv": "2407.04346"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.04346"}], "name": "MobileFlow: A Multimodal LLM For Mobile GUI Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a mobile GUI model using hybrid visual encoders, mixture-of-experts expansion, and alignment training for multilingual interfaces.", "topics": ["mobile-navigation", "screen-understanding"], "updated_at": "2024-12-06", "url": "https://arxiv.org/abs/2407.04346"}
{"aliases": [], "authors": ["Petr Anokhin", "Nikita Semenov", "Artyom Sorokin", "Dmitry Evseev", "Andrey Kravchenko", "Mikhail Burtsev", "Evgeny Burnaev"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4f01004cc70f18c14a24cb22ccda9e58613d1e9e9fbd14551573216beeca5875", "submitted_at": "2024-07-05T09:06:47Z", "updated_at": "2025-05-15T10:57:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.04363v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.04363", "identifiers": {"arxiv": "2407.04363"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.04363"}], "name": "AriGraph: Learning Knowledge Graph World Models with Episodic Memory for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-15", "url": "https://arxiv.org/abs/2407.04363"}
{"aliases": [], "authors": ["Nathan Gavenski", "Juarez Monteiro", "Felipe Meneguzzi", "Michael Luck", "Odinaldo Rodrigues"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "101c1214864fdcd2daeb0b87a60ab43568a538179f913fd5fbb8b0ee71c01d5e", "submitted_at": "2024-07-05T20:25:39Z", "updated_at": "2024-07-22T15:32:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.04856v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.04856", "identifiers": {"arxiv": "2407.04856"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.04856"}], "name": "Explorative Imitation Learning: A Path Signature Approach for Continuous Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-22", "url": "https://arxiv.org/abs/2407.04856"}
{"aliases": [], "authors": ["Léo Boisvert", "Megh Thakkar", "Maxime Gasse", "Massimo Caccia", "Thibault Le Sellier De Chezelles", "Quentin Cappart", "Nicolas Chapados", "Alexandre Lacoste", "Alexandre Drouin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b56a6208e4fa3523cae3fbae673affcd5fe3d6a9c19df482d27fe71d43ffef5", "submitted_at": "2024-07-07T07:15:49Z", "updated_at": "2025-02-05T21:50:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.05291v2"}, "publication": {"authors": ["Léo Boisvert", "Megh Thakkar", "Maxime Gasse", "Massimo Caccia", "Thibault L. De Chezelles", "Quentin Cappart", "Nicolas Chapados", "Alexandre Lacoste", "Alexandre Drouin"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-0195", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "5996-6051", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/0b82662b6c32e887bb252a74d8cb2d5e-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "WorkArena++: Towards Compositional Planning and Reasoning-based Common Knowledge Work Tasks", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:04Z", "role": "primary", "sha256": "4fe01789f93df840699d604de87c068557fc02aa78d22d53c6b3ae7c9db69cc4", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.05291", "identifiers": {"arxiv": "2407.05291"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.05291"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/0b82662b6c32e887bb252a74d8cb2d5e-Paper-Datasets_and_Benchmarks_Track.pdf"}], "name": "WorkArena++: Towards Compositional Planning and Reasoning-based Common Knowledge Work Tasks", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-07-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Defines compositional enterprise workflows to evaluate planning and reasoning, with generation of ground-truth interaction traces.", "topics": ["evaluation", "planning", "workflow"], "updated_at": "2025-02-05", "url": "https://arxiv.org/abs/2407.05291"}
{"aliases": [], "authors": ["Alan Wu", "Ye Yuan", "Zhiping Xiao", "Ming Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cb4140a98aab68aac176e14d6818fa961669b9376db35bc450821b2ca226ec3", "submitted_at": "2024-07-08T15:51:37Z", "updated_at": "2026-04-20T07:55:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.06048v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.06048", "identifiers": {"arxiv": "2407.06048"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.06048"}], "name": "Vision-Braille: A Curriculum Learning Toolkit and Braille-Chinese Corpus for Braille Translation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-20", "url": "https://arxiv.org/abs/2407.06048"}
{"aliases": [], "authors": ["Weize Chen", "Ziming You", "Ran Li", "Yitong Guan", "Chen Qian", "Chenyang Zhao", "Cheng Yang", "Ruobing Xie", "Zhiyuan Liu", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9a33f39a6b7aea08e080646cc6c327ec944f79d3446aecdf4661090687702a90", "submitted_at": "2024-07-09T17:33:24Z", "updated_at": "2024-07-10T15:57:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.07061v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.07061", "identifiers": {"arxiv": "2407.07061"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.07061"}], "name": "Internet of Agents: Weaving a Web of Heterogeneous Agents for Collaborative Intelligence", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-10", "url": "https://arxiv.org/abs/2407.07061"}
{"aliases": [], "authors": ["Peitong Duan", "Chin-yi Chen", "Gang Li", "Bjoern Hartmann", "Yang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "390a7415ad5f665de43f7d294e0a575785ef7e2810dead3363632174b2986af4", "submitted_at": "2024-07-11T20:18:19Z", "updated_at": "2024-08-13T23:41:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.08850v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.08850", "identifiers": {"arxiv": "2407.08850"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.08850"}], "name": "UICrit: Enhancing Automated Design Evaluation with a UICritique Dataset", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-11", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces designer-authored mobile UI critiques and quality ratings, studying how targeted examples improve automated design feedback.", "topics": ["screen-understanding", "evaluation"], "updated_at": "2024-08-13", "url": "https://arxiv.org/abs/2407.08850"}
{"aliases": [], "authors": ["Shrestha Mohanty", "Negar Arabzadeh", "Andrea Tupini", "Yuxuan Sun", "Alexey Skrynnik", "Artem Zholus", "Marc-Alexandre Côté", "Julia Kiseleva"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0510e37d77d7063e8b03d50ec421cb2b5ccfc5ace5efc9f096de7078fd042354", "submitted_at": "2024-07-12T00:07:43Z", "updated_at": "2024-07-12T00:07:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.08898v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.08898", "identifiers": {"arxiv": "2407.08898"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.08898"}], "name": "IDAT: A Multi-Modal Dataset and Toolkit for Building and Evaluating Interactive Task-Solving Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-12", "url": "https://arxiv.org/abs/2407.08898"}
{"aliases": [], "authors": ["Yongxiang Hu", "Xuan Wang", "Yingchuan Wang", "Yu Zhang", "Shiyu Guo", "Chaoyi Chen", "Xin Wang", "Yangfan Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a594fa7e150c9f5c22ed4d88f0551305f801f5780851ecf844802f2d2cd07c48", "submitted_at": "2024-07-12T06:14:46Z", "updated_at": "2024-07-12T06:14:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.09018v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.09018", "identifiers": {"arxiv": "2407.09018"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.09018"}], "name": "AUITestAgent: Automatic Requirements Oriented GUI Function Testing", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Automates mobile GUI tests from natural-language requirements, combining action generation with verification based on interaction traces.", "topics": ["automation", "evaluation"], "updated_at": "2024-07-12", "url": "https://arxiv.org/abs/2407.09018"}
{"aliases": [], "authors": ["Haoyu Dong", "Jianbo Zhao", "Yuzhang Tian", "Junyu Xiong", "Shiyu Xia", "Mengyu Zhou", "Yun Lin", "José Cambronero", "Yeye He", "Shi Han", "Dongmei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6acb24dcbd3cfc1745c513b82d01b62c7fded6931353da50b76ca9052adbc561", "submitted_at": "2024-07-12T06:34:21Z", "updated_at": "2025-04-02T14:33:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.09025v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.09025", "identifiers": {"arxiv": "2407.09025"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.09025"}], "name": "SpreadsheetLLM: Encoding Spreadsheets for Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-02", "url": "https://arxiv.org/abs/2407.09025"}
{"aliases": [], "authors": ["Yulong Yang", "Xinshan Yang", "Shuaidong Li", "Chenhao Lin", "Zhengyu Zhao", "Chao Shen", "Tianwei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88576325385d3c361f0666311eae8858ea0d4a0098d73221f789550bebf1e5db", "submitted_at": "2024-07-12T14:30:05Z", "updated_at": "2025-03-16T07:13:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.09295v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2407.09295", "identifiers": {"arxiv": "2407.09295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.09295"}], "name": "Systematic Categorization, Construction and Evaluation of New Attacks against Multi-modal Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Develops a threat taxonomy and evaluation framework for attacks against multimodal mobile GUI agents, with case studies and dataset experiments.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-03-16", "url": "https://arxiv.org/abs/2407.09295"}
{"aliases": [], "authors": ["A. Saxena", "M. Salvato", "W. Roster", "R. Shirley", "J. Buchner", "J. Wolf", "C. Kohl", "H. Starck", "T. Dwelly", "J. Comparat", "A. Malyali", "S. Krippendorf", "A. Zenteno", "D. Lang", "D. Schlegel", "R. Zhou", "A. Dey", "F. Valdes", "A. Myers", "R. J. Assef", "C. Ricci", "M. J. Temple", "A. Merloni", "A. Koekemoer", "S. F. Anderson", "S. Morrison", "X. Liu", "K. Nandra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ee6439faa2c13d241edaec8a26e693694ec3720436b173c34580fb892ce5e739", "submitted_at": "2024-07-15T15:07:30Z", "updated_at": "2024-07-20T15:38:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.10788v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2407.10788", "identifiers": {"arxiv": "2407.10788"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.10788"}], "name": "CircleZ: Reliable Photometric redshifts for AGN computed using only photometry from Legacy Survey Imaging for DESI", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-20", "url": "https://arxiv.org/abs/2407.10788"}
{"aliases": [], "authors": ["Carlo Romeo", "Andrew D. Bagdanov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "80e072ec1774e3caff1b18cc6112afb976c08cd8f83c4736a2d0efa7616d3ce5", "submitted_at": "2024-07-15T15:53:13Z", "updated_at": "2024-07-15T15:53:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.10839v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.10839", "identifiers": {"arxiv": "2407.10839"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.10839"}], "name": "Offline Reinforcement Learning with Imputed Rewards", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-15", "url": "https://arxiv.org/abs/2407.10839"}
{"aliases": [], "authors": ["Ruisheng Cao", "Fangyu Lei", "Haoyuan Wu", "Jixuan Chen", "Yeqiao Fu", "Hongcheng Gao", "Xinzhuang Xiong", "Hanchong Zhang", "Yuchen Mao", "Wenjing Hu", "Tianbao Xie", "Hongshen Xu", "Danyang Zhang", "Sida Wang", "Ruoxi Sun", "Pengcheng Yin", "Caiming Xiong", "Ansong Ni", "Qian Liu", "Victor Zhong", "Lu Chen", "Kai Yu", "Tao Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7922a723142ac62018064fe5fc9357fb77c08e316ac4d6d6ba728f1b337df690", "submitted_at": "2024-07-15T17:54:37Z", "updated_at": "2024-07-15T17:54:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.10956v1"}, "publication": {"authors": ["Ruisheng Cao", "Fangyu Lei", "Haoyuan Wu", "Jixuan Chen", "Yeqiao Fu", "Hongcheng Gao", "Xinzhuang Xiong", "Hanchong Zhang", "Yuchen Mao", "Wenjing Hu", "Tianbao Xie", "Hongshen Xu", "Danyang Zhang", "Sida Wang", "Ruoxi Sun", "Pengcheng Yin", "Caiming Xiong", "Ansong Ni", "Qian Liu", "Victor Zhong", "Lu Chen", "Kai Yu", "Tao Yu"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-3421", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "107703-107744", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/c2f71567cd53464161cab3336e8fc865-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "Spider2-V: How Far Are Multimodal Agents From Automating Data Science and Engineering Workflows?", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:04Z", "role": "primary", "sha256": "872286a0210616be9737de716cc6069a1e9cbee1df3182d45c718582e185f62f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.10956", "identifiers": {"arxiv": "2407.10956"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.10956"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/c2f71567cd53464161cab3336e8fc865-Paper-Datasets_and_Benchmarks_Track.pdf"}], "name": "Spider2-V: How Far Are Multimodal Agents From Automating Data Science and Engineering Workflows?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Benchmarks multimodal agents on data-science and engineering workflows involving enterprise applications, code, and GUI operations in real computer environments.", "topics": ["evaluation", "science", "workflow"], "updated_at": "2024-07-15", "url": "https://arxiv.org/abs/2407.10956"}
{"aliases": [], "authors": ["Zelong Li", "Shuyuan Xu", "Kai Mei", "Wenyue Hua", "Balaji Rama", "Om Raheja", "Hao Wang", "He Zhu", "Yongfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7312bcd35ac6a90f4c420db74f2d9ae779a0a731659b4d0c471820397795da68", "submitted_at": "2024-07-01T21:05:02Z", "updated_at": "2024-07-01T21:05:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.12821v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.12821", "identifiers": {"arxiv": "2407.12821"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.12821"}], "name": "AutoFlow: Automated Workflow Generation for Large Language Model Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-01", "url": "https://arxiv.org/abs/2407.12821"}
{"aliases": [], "authors": ["Tamer Abuelsaad", "Deepak Akkil", "Prasenjit Dey", "Ashish Jagmohan", "Aditya Vempaty", "Ravi Kokku"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "18cfd29fcb66d703d90f747c162db1b3fc6a43f400b4eacbddf081bb83ed7947", "submitted_at": "2024-07-17T21:44:28Z", "updated_at": "2024-07-17T21:44:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.13032v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.13032", "identifiers": {"arxiv": "2407.13032"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.13032"}], "name": "Agent-E: From Autonomous Web Navigation to Foundational Design Principles in Agentic Systems", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Agent-E combines hierarchical web-task control with DOM distillation and observations of interface changes. WebVoyager experiments inform a broader discussion of primitive skills, observation reduction, and self-improvement in agent architecture.", "topics": ["web-navigation", "planning", "efficiency", "state-tracking"], "updated_at": "2024-07-17", "url": "https://arxiv.org/abs/2407.13032"}
{"aliases": [], "authors": ["Mokter M. Chowdhury", "Robinson Cortes-Huerto", "Debashish Mukherji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5151732ea57dac1d7e4ad730d5cf9c53caa8ba97035645b8509ef4e211171ac", "submitted_at": "2024-07-18T00:13:12Z", "updated_at": "2024-07-18T00:13:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.13063v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.13063", "identifiers": {"arxiv": "2407.13063"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.13063"}], "name": "Smart polymer solution and thermal conductivity: How important is an exact polymer conformation?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-18", "url": "https://arxiv.org/abs/2407.13063"}
{"aliases": [], "authors": ["Yotam Perlitz", "Ariel Gera", "Ofir Arviv", "Asaf Yehudai", "Elron Bandel", "Eyal Shnarch", "Michal Shmueli-Scheuer", "Leshem Choshen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0380ae6377082f916dbfa5a5d81b6679cf35159551c80e98e308096c5fc1b531", "submitted_at": "2024-07-18T17:00:23Z", "updated_at": "2024-09-12T08:36:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.13696v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.13696", "identifiers": {"arxiv": "2407.13696"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.13696"}], "name": "Do These LLM Benchmarks Agree? Fixing Benchmark Evaluation with BenchBench", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-12", "url": "https://arxiv.org/abs/2407.13696"}
{"aliases": [], "authors": ["Shunyu Liu", "Yaoru Li", "Kongcheng Zhang", "Zhenyu Cui", "Wenkai Fang", "Yuxuan Zheng", "Tongya Zheng", "Mingli Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d19cb1092f10a9dbd3a9326a03391bb82275cdc1fd5debb2af3fa8e432cac340", "submitted_at": "2024-07-22T02:06:59Z", "updated_at": "2025-05-31T06:58:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.15325v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2407.15325", "identifiers": {"arxiv": "2407.15325"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.15325"}], "name": "Odyssey: Empowering Minecraft Agents with Open-World Skills", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-31", "url": "https://arxiv.org/abs/2407.15325"}
{"aliases": [], "authors": ["Segev Shlomov", "Avi Yaeli", "Sami Marreed", "Sivan Schwartz", "Netanel Eder", "Offer Akrabi", "Sergey Zeltyn"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4caa87e0b88d9375a23ca412ed19b3806f67712ca4ac222a39746e08ee80b07d", "submitted_at": "2024-07-22T14:41:32Z", "updated_at": "2024-08-18T09:02:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.15673v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.15673", "identifiers": {"arxiv": "2407.15673"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.15673"}], "name": "IDA: Breaking Barriers in No-code UI Automation Through Large Language Models and Human-Centric Design", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "IDA helps business users create web UI automation through guided demonstrations, a semantic programming model, and a teaching-oriented interface. A prototype user study examines automation creation with enterprise applications by users without programming backgrounds.", "topics": ["human-in-the-loop", "automation", "workflow", "industry"], "updated_at": "2024-08-18", "url": "https://arxiv.org/abs/2407.15673"}
{"aliases": [], "authors": ["Ori Yoran", "Samuel Joseph Amouyal", "Chaitanya Malaviya", "Ben Bogin", "Ofir Press", "Jonathan Berant"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f028fb7f4d64fd4c29886d49b5d6f0fabddba309b669e019edf13586deb9c90", "submitted_at": "2024-07-22T15:18:45Z", "updated_at": "2024-10-21T15:45:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.15711v2"}, "publication": {"authors": ["Ori Yoran", "Samuel Joseph Amouyal", "Chaitanya Malaviya", "Ben Bogin", "Ofir Press", "Jonathan Berant"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.emnlp-main.505", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.emnlp-main.505/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "8938-8968", "pdf_url": "https://aclanthology.org/2024.emnlp-main.505.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AssistantBench: Can Web Agents Solve Realistic and Time-Consuming Tasks?", "type": "conference-paper", "url": "https://aclanthology.org/2024.emnlp-main.505/", "venue": "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "a18f0ea465d83796bcd83b95f48c3c13cbdc0aa778ea0dd454aa61f8082920b4", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.emnlp-main.505/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2407.15711", "identifiers": {"arxiv": "2407.15711"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.15711"}, {"rel": "publication", "url": "https://aclanthology.org/2024.emnlp-main.505/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.emnlp-main.505.pdf"}], "name": "AssistantBench: Can Web Agents Solve Realistic and Time-Consuming Tasks?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces realistic open-web tasks and the SeePlanAct agent, analyzing the limitations of browsing and retrieval in time-consuming assistant work.", "topics": ["evaluation", "web-navigation", "planning"], "updated_at": "2024-10-21", "url": "https://arxiv.org/abs/2407.15711"}
{"aliases": [], "authors": ["Xingyao Wang", "Boxuan Li", "Yufan Song", "Frank F. Xu", "Xiangru Tang", "Mingchen Zhuge", "Jiayi Pan", "Yueqi Song", "Bowen Li", "Jaskirat Singh", "Hoang H. Tran", "Fuqiang Li", "Ren Ma", "Mingzhang Zheng", "Bill Qian", "Yanjun Shao", "Niklas Muennighoff", "Yizhe Zhang", "Binyuan Hui", "Junyang Lin", "Robert Brennan", "Hao Peng", "Heng Ji", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cb94bd5d52d4abf7ff8511a64ac49d4699c795baa0446cd47e2b79616cc10408", "submitted_at": "2024-07-23T17:50:43Z", "updated_at": "2025-04-18T18:14:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.16741v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.16741", "identifiers": {"arxiv": "2407.16741"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.16741"}], "name": "OpenHands: An Open Platform for AI Software Developers as Generalist Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-18", "url": "https://arxiv.org/abs/2407.16741"}
{"aliases": [], "authors": ["Yuxiang Chai", "Siyuan Huang", "Yazhe Niu", "Han Xiao", "Liang Liu", "Dingyu Zhang", "Shuai Ren", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c80665716bf665fe4688a27a4c7dcf26fcfe93cb7f79a6f28784a795e0bd3cb", "submitted_at": "2024-07-03T17:59:58Z", "updated_at": "2025-05-29T02:43:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.17490v2"}, "publication": {"authors": ["Yuxiang Chai", "Siyuan Huang", "Yazhe Niu", "Han Xiao", "Liang Liu", "Guozhi Wang", "Dingyu Zhang", "Shuai Ren", "Hongsheng Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.110", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.110/", "https://aclanthology.org/2025.findings-acl.110.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. The publisher PDF, page 1, lists nine authors including Guozhi Wang; the retained arXiv snapshot lists eight. Each edition retains its own authors.", "pages": "2138-2156", "pdf_url": "https://aclanthology.org/2025.findings-acl.110.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AMEX: Android Multi-annotation Expo Dataset for Mobile GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.110/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:52Z", "role": "primary", "sha256": "9cb5bd38791ee3df95d04815ca650bf0f020518b009c04b4a54857c122f161c8", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.110/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:02.000709+00:00", "role": "primary", "sha256": "344342e1c229e46269d26d879b32a4d2947932607b968f7e18382d231ad05ee2", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2025.findings-acl.110.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2407.17490", "identifiers": {"arxiv": "2407.17490"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.17490"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.110/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.110.pdf"}], "name": "AMEX: Android Multi-annotation Expo Dataset for Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Android screenshots with element locations, functionality descriptions, and instructions linked to sequences of GUI actions.", "topics": ["grounding", "screen-understanding", "mobile-navigation"], "updated_at": "2025-05-29", "url": "https://arxiv.org/abs/2407.17490"}
{"aliases": [], "authors": ["Xuchen Pan", "Dawei Gao", "Yuexiang Xie", "Yushuo Chen", "Zhewei Wei", "Yaliang Li", "Bolin Ding", "Ji-Rong Wen", "Jingren Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "beed2f478ec95592641378b6e3a4fd14dcd306612c042bf10bd4ccafc6266933", "submitted_at": "2024-07-25T05:50:46Z", "updated_at": "2024-10-28T13:19:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.17789v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.17789", "identifiers": {"arxiv": "2407.17789"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.17789"}], "name": "Very Large-Scale Multi-Agent Simulation in AgentScope", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-28", "url": "https://arxiv.org/abs/2407.17789"}
{"aliases": [], "authors": ["Harsh Trivedi", "Tushar Khot", "Mareike Hartmann", "Ruskin Manku", "Vinty Dong", "Edward Li", "Shashank Gupta", "Ashish Sabharwal", "Niranjan Balasubramanian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e59087ecf28d07baf81b60e4166a866d4bea4d8cd934bea06c2149dd61f38f1a", "submitted_at": "2024-07-26T17:55:45Z", "updated_at": "2024-07-26T17:55:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.18901v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.18901", "identifiers": {"arxiv": "2407.18901"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.18901"}], "name": "AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-26", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces simulated application APIs and state-based task evaluation for interactive coding agents performing everyday digital workflows.", "topics": ["evaluation", "tool-integration", "cross-app"], "updated_at": "2024-07-26", "url": "https://arxiv.org/abs/2407.18901"}
{"aliases": [], "authors": ["Zilong Wang", "Yuedong Cui", "Li Zhong", "Zimin Zhang", "Da Yin", "Bill Yuchen Lin", "Jingbo Shang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c443e6a508c438332ac5990cf254a4441f5bdcc24354fd8aeca36bcbf4604a29", "submitted_at": "2024-07-26T19:27:17Z", "updated_at": "2024-07-26T19:27:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.19056v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.19056", "identifiers": {"arxiv": "2407.19056"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.19056"}], "name": "OfficeBench: Benchmarking Language Agents across Multiple Applications for Office Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-07-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates language agents on office workflows requiring information integration, long-term planning, and switching between applications.", "topics": ["evaluation", "cross-app", "long-horizon"], "updated_at": "2024-07-26", "url": "https://arxiv.org/abs/2407.19056"}
{"aliases": [], "authors": ["Borui Wan", "Mingji Han", "Yiyao Sheng", "Yanghua Peng", "Haibin Lin", "Mofan Zhang", "Zhichao Lai", "Menghan Yu", "Junda Zhang", "Zuquan Song", "Xin Liu", "Chuan Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "930c7c85f4bbd72fd5c06b9ee196a53f2ce7e4f051a75cfa81b3bda324304032", "submitted_at": "2024-07-29T16:18:20Z", "updated_at": "2025-04-02T06:05:23Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2407.20143v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.20143", "identifiers": {"arxiv": "2407.20143"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.20143"}], "name": "ByteCheckpoint: A Unified Checkpointing System for Large Foundation Model Development", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-02", "url": "https://arxiv.org/abs/2407.20143"}
{"aliases": [], "authors": ["Zehui Chen", "Kuikun Liu", "Qiuchen Wang", "Jiangning Liu", "Wenwei Zhang", "Kai Chen", "Feng Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "652a6381d2a6f960a2775168711be531e4994daeca95bc022716a2babae7c2ed", "submitted_at": "2024-07-29T17:12:40Z", "updated_at": "2025-10-31T02:40:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.20183v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2407.20183", "identifiers": {"arxiv": "2407.20183"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.20183"}], "name": "MindSearch: Mimicking Human Minds Elicits Deep AI Searcher", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-31", "url": "https://arxiv.org/abs/2407.20183"}
{"aliases": [], "authors": ["Başak Melis Öcal", "Maxim Tatarchenko", "Sezer Karaoglu", "Theo Gevers"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a01440dded23d610f757f8bd88e376c0d3d76481972ee5c4047ebf39aef19067", "submitted_at": "2024-07-30T10:45:28Z", "updated_at": "2024-07-30T10:45:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.20727v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2407.20727", "identifiers": {"arxiv": "2407.20727"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.20727"}], "name": "SceneTeller: Language-to-3D Scene Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-07-30", "url": "https://arxiv.org/abs/2407.20727"}
{"aliases": [], "authors": ["Tom Gunter", "Zirui Wang", "Chong Wang", "Ruoming Pang", "Andy Narayanan", "Aonan Zhang", "Bowen Zhang", "Chen Chen", "Chung-Cheng Chiu", "David Qiu", "Deepak Gopinath", "Dian Ang Yap", "Dong Yin", "Feng Nan", "Floris Weers", "Guoli Yin", "Haoshuo Huang", "Jianyu Wang", "Jiarui Lu", "John Peebles", "Ke Ye", "Mark Lee", "Nan Du", "Qibin Chen", "Quentin Keunebroek", "Sam Wiseman", "Syd Evans", "Tao Lei", "Vivek Rathod", "Xiang Kong", "Xianzhi Du", "Yanghao Li", "Yongqiang Wang", "Yuan Gao", "Zaid Ahmed", "Zhaoyang Xu", "Zhiyun Lu", "Al Rashid", "Albin Madappally Jose", "Alec Doane", "Alfredo Bencomo", "Allison Vanderby", "Andrew Hansen", "Ankur Jain", "Anupama Mann Anupama", "Areeba Kamal", "Bugu Wu", "Carolina Brum", "Charlie Maalouf", "Chinguun Erdenebileg", "Chris Dulhanty", "Daniel Parilla", "Dominik Moritz", "Doug Kang", "Eduardo Jimenez", "Evan Ladd", "Fangping Shi", "Felix Bai", "Frank Chu", "Fred Hohman", "Hadas Kotek", "Hannah Gillis Coleman", "Jane Li", "Jeffrey Bigham", "Jeffery Cao", "Jeff Lai", "Jessica Cheung", "Jiulong Shan", "Joe Zhou", "John Li", "Jun Qin", "Karanjeet Singh", "Karla Vega", "Kelvin Zou", "Laura Heckman", "Lauren Gardiner", "Margit Bowler", "Maria Cordell", "Meng Cao", "Nicole Hay", "Nilesh Shahdadpuri", "Otto Godwin", "Pranay Dighe", "Pushyami Rachapudi", "Ramsey Tantawi", "Roman Frigg", "Sam Davarnia", "Sanskruti Shah", "Saptarshi Guha", "Sasha Sirovica", "Shen Ma", "Shuang Ma", "Simon Wang", "Sulgi Kim", "Suma Jayaram", "Vaishaal Shankar", "Varsha Paidi", "Vivek Kumar", "Xin Wang", "Xin Zheng", "Walker Cheng", "Yael Shrager", "Yang Ye", "Yasu Tanaka", "Yihao Guo", "Yunsong Meng", "Zhao Tang Luo", "Zhi Ouyang", "Alp Aygar", "Alvin Wan", "Andrew Walkingshaw", "Andy Narayanan", "Antonie Lin", "Arsalan Farooq", "Brent Ramerth", "Colorado Reed", "Chris Bartels", "Chris Chaney", "David Riazati", "Eric Liang Yang", "Erin Feldman", "Gabriel Hochstrasser", "Guillaume Seguin", "Irina Belousova", "Joris Pelemans", "Karen Yang", "Keivan Alizadeh Vahid", "Liangliang Cao", "Mahyar Najibi", "Marco Zuliani", "Max Horton", "Minsik Cho", "Nikhil Bhendawade", "Patrick Dong", "Piotr Maj", "Pulkit Agrawal", "Qi Shan", "Qichen Fu", "Regan Poston", "Sam Xu", "Shuangning Liu", "Sushma Rao", "Tashweena Heeramun", "Thomas Merth", "Uday Rayala", "Victor Cui", "Vivek Rangarajan Sridhar", "Wencong Zhang", "Wenqi Zhang", "Wentao Wu", "Xingyu Zhou", "Xinwen Liu", "Yang Zhao", "Yin Xia", "Zhile Ren", "Zhongzheng Ren"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "668c2f6a4b32878be680517529053846300bfc4f8bd511abd465764efa4478b8", "submitted_at": "2024-07-29T18:38:49Z", "updated_at": "2026-05-27T04:08:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.21075v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2407.21075", "identifiers": {"arxiv": "2407.21075"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2407.21075"}], "name": "Apple Intelligence Foundation Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2407.21075"}
{"aliases": [], "authors": ["Yadong Lu", "Jianwei Yang", "Yelong Shen", "Ahmed Awadallah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfe55dd83d2f5ac48f9e3efb6bdf4aab7f11d81c0118725555456cce276adae3", "submitted_at": "2024-08-01T00:00:43Z", "updated_at": "2024-08-01T00:00:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.00203v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2408.00203", "identifiers": {"arxiv": "2408.00203"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.00203"}], "name": "OmniParser for Pure Vision Based GUI Agent", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-08-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Parses screenshots into interactable regions and semantic element descriptions to support visually grounded agent actions.", "topics": ["grounding", "screen-understanding"], "updated_at": "2024-08-01", "url": "https://arxiv.org/abs/2408.00203"}
{"aliases": [], "authors": ["Yuxin Lin", "Jiaxuan Ma", "Sizhe Gu", "Jipeng Kong", "Bowen Xu", "Xiting Zhao", "Dengji Zhao", "Wenhan Cao", "Sören Schwertfeger"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a39dc02eb5261815e2100b5277e86d9f26f6d769cac28bca032739fde54ddf4c", "submitted_at": "2024-08-01T13:29:26Z", "updated_at": "2024-08-01T13:29:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.00545v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2408.00545", "identifiers": {"arxiv": "2408.00545"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.00545"}], "name": "Collecting Larg-Scale Robotic Datasets on a High-Speed Mobile Platform", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-01", "url": "https://arxiv.org/abs/2408.00545"}
{"aliases": [], "authors": ["Ali Al-Bayaty", "Xiaoyu Song", "Marek Perkowski"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c8f8ef90c35eab96bcadd20478e85ba850e691d9db6f3671849fbad15492352b", "submitted_at": "2024-08-02T05:50:35Z", "updated_at": "2024-08-02T05:50:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.01025v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2408.01025", "identifiers": {"arxiv": "2408.01025"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.01025"}], "name": "CALA-$n$: A Quantum Library for Realizing Cost-Effective 2-, 3-, 4-, and 5-bit Gates on IBM Quantum Computers using Bloch Sphere Approach, Clifford+T Gates, and Layouts", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-02", "url": "https://arxiv.org/abs/2408.01025"}
{"aliases": [], "authors": ["Andrew Zhu", "Liam Dugan", "Chris Callison-Burch"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bc30213f187e4399379015695f9476746d22a900fff72090fba1c1f39e827107", "submitted_at": "2024-08-05T05:43:23Z", "updated_at": "2024-11-04T19:16:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2408.02248v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2408.02248", "identifiers": {"arxiv": "2408.02248"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.02248"}], "name": "ReDel: A Toolkit for LLM-Powered Recursive Multi-Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-04", "url": "https://arxiv.org/abs/2408.02248"}
{"aliases": [], "authors": ["Xinbei Ma", "Yiting Wang", "Yao Yao", "Tongxin Yuan", "Aston Zhang", "Zhuosheng Zhang", "Hai Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d66c0c2102dd2734d5334faeb12b4971d50a2d2e50a6671ede2fa78a84a0c03f", "submitted_at": "2024-08-05T15:16:22Z", "updated_at": "2025-09-05T09:21:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2408.02544v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2408.02544", "identifiers": {"arxiv": "2408.02544"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.02544"}], "name": "Caution for the Environment: Multimodal LLM Agents are Susceptible to Environmental Distractions", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The abstract describes a simulated dataset; real-browser deployment is not asserted.", "published_at": "2024-08-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Measures how irrelevant GUI content affects multimodal agents using a simulated dataset and three perception patterns. The study also examines adversarial environment injection and possible ways to improve instruction faithfulness.", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2025-09-05", "url": "https://arxiv.org/abs/2408.02544"}
{"aliases": [], "authors": ["Bo Li", "Yuanhan Zhang", "Dong Guo", "Renrui Zhang", "Feng Li", "Hao Zhang", "Kaichen Zhang", "Peiyuan Zhang", "Yanwei Li", "Ziwei Liu", "Chunyuan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "001c23a98b9b9799e1c3c5d7dd6a886380c947ceae326697ce230325afad31eb", "submitted_at": "2024-08-06T17:59:44Z", "updated_at": "2024-10-26T16:35:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2408.03326v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2408.03326", "identifiers": {"arxiv": "2408.03326"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.03326"}], "name": "LLaVA-OneVision: Easy Visual Task Transfer", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-26", "url": "https://arxiv.org/abs/2408.03326"}
{"aliases": [], "authors": ["Qingbin Zeng", "Qinglong Yang", "Shunan Dong", "Heming Du", "Liang Zheng", "Fengli Xu", "Yong Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3442e96754d426bb1663fb01effdbf7532489bff961d20d7aa009abb44018a8e", "submitted_at": "2024-08-08T02:28:43Z", "updated_at": "2024-10-17T06:43:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2408.04168v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2408.04168", "identifiers": {"arxiv": "2408.04168"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.04168"}], "name": "Perceive, Reflect, and Plan: Designing LLM Agent for Goal-Directed City Navigation without Instructions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-17", "url": "https://arxiv.org/abs/2408.04168"}
{"aliases": [], "authors": ["Xiao Liu", "Tianjie Zhang", "Yu Gu", "Iat Long Iong", "Yifan Xu", "Xixuan Song", "Shudan Zhang", "Hanyu Lai", "Xinyi Liu", "Hanlin Zhao", "Jiadai Sun", "Xinyue Yang", "Yu Yang", "Zehan Qi", "Shuntian Yao", "Xueqiao Sun", "Siyi Cheng", "Qinkai Zheng", "Hao Yu", "Hanchen Zhang", "Wenyi Hong", "Ming Ding", "Lihang Pan", "Xiaotao Gu", "Aohan Zeng", "Zhengxiao Du", "Chan Hee Song", "Yu Su", "Yuxiao Dong", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "603bfa4357617bfeb7a718065e77c71d77614f7240258684938dd4e6fd1a20df", "submitted_at": "2024-08-12T17:44:17Z", "updated_at": "2024-08-12T17:44:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.06327v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2408.06327", "identifiers": {"arxiv": "2408.06327"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.06327"}], "name": "VisualAgentBench: Towards Large Multimodal Models as Visual Foundation Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-08-12", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Benchmarks visual agents across GUI, embodied, and design tasks, with training trajectories collected through solvers, agents, and human demonstrations.", "topics": ["evaluation", "data-generation"], "updated_at": "2024-08-12", "url": "https://arxiv.org/abs/2408.06327"}
{"aliases": [], "authors": ["Pranav Putta", "Edmund Mills", "Naman Garg", "Sumeet Motwani", "Chelsea Finn", "Divyansh Garg", "Rafael Rafailov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d7171ec4f2a16cb7c0416172e93b04197aac992ec0d324bfbc13ddae8253450c", "submitted_at": "2024-08-13T20:52:13Z", "updated_at": "2024-08-13T20:52:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.07199v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2408.07199", "identifiers": {"arxiv": "2408.07199"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.07199"}], "name": "Agent Q: Advanced Reasoning and Learning for Autonomous AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-13", "url": "https://arxiv.org/abs/2408.07199"}
{"aliases": [], "authors": ["Alessandro Berti", "Mayssa Maatallah", "Urszula Jessen", "Michal Sroka", "Sonia Ayachi Ghannouchi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f7a4aa11affbaddcfca0a833244b8fc861632f5435af81d36ac3200db701acfe", "submitted_at": "2024-08-14T10:14:18Z", "updated_at": "2024-08-14T10:14:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.07720v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2408.07720", "identifiers": {"arxiv": "2408.07720"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.07720"}], "name": "Re-Thinking Process Mining in the AI-Based Agents Era", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-14", "url": "https://arxiv.org/abs/2408.07720"}
{"aliases": [], "authors": ["Mengkang Hu", "Tianxing Chen", "Qiguang Chen", "Yao Mu", "Wenqi Shao", "Ping Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "30662f6dc262461b3f2469763b1abdf4320b7ceb2e7c7e01e525eccd92271bb6", "submitted_at": "2024-08-18T17:59:49Z", "updated_at": "2024-08-18T17:59:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.09559v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2408.09559", "identifiers": {"arxiv": "2408.09559"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.09559"}], "name": "HiAgent: Hierarchical Working Memory Management for Solving Long-Horizon Agent Tasks with Large Language Model", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-18", "url": "https://arxiv.org/abs/2408.09559"}
{"aliases": [], "authors": ["Yuki Kurokawa", "Junichi Takahashi", "Yoshiya Yamanaka"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29331803949aa69fb36927d21b66967cbe062859319c7aded2a2c83de8874ac6", "submitted_at": "2024-08-18T19:34:24Z", "updated_at": "2024-08-22T13:27:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2408.09577v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2408.09577", "identifiers": {"arxiv": "2408.09577"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.09577"}], "name": "Quantum Reservoir Computing Using Bose-Einstein Condensate with Damping", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-22", "url": "https://arxiv.org/abs/2408.09577"}
{"aliases": [], "authors": ["Sinan Ibrahim", "Mostafa Mostafa", "Ali Jnadi", "Hadi Salloum", "Pavel Osinenko"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "505e6d795d5e3c1f6837351f66e39b4d1b7bda0c54607837e978ce6bed3c75e2", "submitted_at": "2024-07-22T09:28:12Z", "updated_at": "2024-12-27T19:25:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2408.10215v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2408.10215", "identifiers": {"arxiv": "2408.10215"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.10215"}], "name": "Comprehensive Overview of Reward Engineering and Shaping in Advancing Reinforcement Learning Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-07-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-27", "url": "https://arxiv.org/abs/2408.10215"}
{"aliases": [], "authors": ["Yanbo Ding", "Shaobin Zhuang", "Kunchang Li", "Zhengrong Yue", "Yu Qiao", "Yali Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "088dd48316e7373fa36d14752606f6f0ec291bab9de64def95e3ca3eadc683e6", "submitted_at": "2024-08-20T07:37:23Z", "updated_at": "2024-12-16T02:08:29Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2408.10605v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2408.10605", "identifiers": {"arxiv": "2408.10605"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.10605"}], "name": "MUSES: 3D-Controllable Image Generation via Multi-Modal Agent Collaboration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-16", "url": "https://arxiv.org/abs/2408.10605"}
{"aliases": [], "authors": ["Patrik Okanovic", "Grzegorz Kwasniewski", "Paolo Sylos Labini", "Maciej Besta", "Flavio Vella", "Torsten Hoefler"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "455cc3de1e2f38ec7a5f13118c67277c35efa58702cb07879dc73ec143918f0c", "submitted_at": "2024-08-21T11:57:42Z", "updated_at": "2024-08-21T11:57:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.11551v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2408.11551", "identifiers": {"arxiv": "2408.11551"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.11551"}], "name": "High Performance Unstructured SpMM Computation Using Tensor Cores", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-21", "url": "https://arxiv.org/abs/2408.11551"}
{"aliases": [], "authors": ["Zachary Ankner", "Mansheej Paul", "Brandon Cui", "Jonathan D. Chang", "Prithviraj Ammanabrolu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cef680aa5d76ba30e376178d1ddf8a345410b61270a4bff9e4e2684fe93fd977", "submitted_at": "2024-08-21T17:24:15Z", "updated_at": "2024-08-21T17:24:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.11791v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2408.11791", "identifiers": {"arxiv": "2408.11791"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.11791"}], "name": "Critique-out-Loud Reward Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-08-21", "url": "https://arxiv.org/abs/2408.11791"}
{"aliases": [], "authors": ["Yanda Li", "Chi Zhang", "Wenjia Jiang", "Wanqi Yang", "Bin Fu", "Pei Cheng", "Xin Chen", "Ling Chen", "Yunchao Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5b471c61c2ac9d4932bd71052ea35585699d70e9710f8366bc4fa8e4f2ebbae", "submitted_at": "2024-08-05T06:31:39Z", "updated_at": "2025-09-17T09:06:58Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2408.11824v4"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2408.11824", "identifiers": {"arxiv": "2408.11824"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.11824"}], "name": "AppAgent v2: Advanced Agent for Flexible Mobile Interactions", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-08-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates exploration and deployment while using a structured application knowledge base and retrieval to guide mobile workflows.", "topics": ["mobile-navigation", "memory"], "updated_at": "2025-09-17", "url": "https://arxiv.org/abs/2408.11824"}
{"aliases": [], "authors": ["Yao Zhang", "Zijian Ma", "Yunpu Ma", "Zhen Han", "Yu Wu", "Volker Tresp"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e3a34f16101df7a96ffdabbdf86b060e195494f16499f20bdf54110a0394cae", "submitted_at": "2024-08-28T17:49:29Z", "updated_at": "2024-08-28T17:49:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.15978v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2408.15978", "identifiers": {"arxiv": "2408.15978"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.15978"}], "name": "WebPilot: A Versatile and Autonomous Multi-Agent System for Web Task Execution with Strategic Exploration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-08-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebPilot separates global subtask planning from local Monte Carlo tree search during web execution. It continually refines the high-level plan and adapts local search to uncertain transitions and incomplete observations, evaluated on WebArena and MiniWoB++.", "topics": ["planning", "web-navigation", "long-horizon"], "updated_at": "2024-08-28", "url": "https://arxiv.org/abs/2408.15978"}
{"aliases": [], "authors": ["Chenglong Wang", "Bongshin Lee", "Steven Drucker", "Dan Marshall", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "629bdc799882be093f8850e4031a69551278978a884fbd105ebcc9e739738cc3", "submitted_at": "2024-08-28T20:12:17Z", "updated_at": "2025-02-21T00:50:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2408.16119v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2408.16119", "identifiers": {"arxiv": "2408.16119"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.16119"}], "name": "Data Formulator 2: Iterative Creation of Data Visualizations, with AI Transforming Data Along the Way", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-21", "url": "https://arxiv.org/abs/2408.16119"}
{"aliases": [], "authors": ["Kumar Gautam"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95aef43a349499537f8e8483417d8057c66a28a92bdb7a2fea57b45dbf658ac7", "submitted_at": "2024-08-30T06:08:35Z", "updated_at": "2024-09-02T05:54:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2408.17035v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2408.17035", "identifiers": {"arxiv": "2408.17035"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2408.17035"}], "name": "Study And Implementation of Unitary Gates in Quantum Computation Using Schrodinger Dynamics", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-02", "url": "https://arxiv.org/abs/2408.17035"}
{"aliases": [], "authors": ["Jiajun Xu", "Zhiyuan Li", "Wei Chen", "Qun Wang", "Xin Gao", "Qi Cai", "Ziyuan Ling"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5aec9e0fa7afad1a6d35e79a7de266eb392e78cc001fb734e68b9d0256e0fa13", "submitted_at": "2024-08-26T03:33:36Z", "updated_at": "2024-09-14T04:01:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.00088v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2409.00088", "identifiers": {"arxiv": "2409.00088"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.00088"}], "name": "On-Device Language Models: A Comprehensive Review", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-14", "url": "https://arxiv.org/abs/2409.00088"}
{"aliases": [], "authors": ["Yijia Shao", "Tianshi Li", "Weiyan Shi", "Yanchen Liu", "Diyi Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d01fedde1aac1a0cf030291d3e152a586db8336dcb1a031e5786ca4af7a625fd", "submitted_at": "2024-08-29T17:58:38Z", "updated_at": "2025-03-14T06:03:20Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.00138v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2409.00138", "identifiers": {"arxiv": "2409.00138"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.00138"}], "name": "PrivacyLens: Evaluating Privacy Norm Awareness of Language Models in Action", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-08-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-14", "url": "https://arxiv.org/abs/2409.00138"}
{"aliases": [], "authors": ["Lutfi Eren Erdogan", "Nicholas Lee", "Siddharth Jha", "Sehoon Kim", "Ryan Tabrizi", "Suhong Moon", "Coleman Hooper", "Gopala Anumanchipalli", "Kurt Keutzer", "Amir Gholami"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f2b4742cf4fe74986bf2c240969a2b9fcb3fdfcb795593b1d4b92adbc2547f1", "submitted_at": "2024-09-01T04:23:48Z", "updated_at": "2024-10-25T01:16:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.00608v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2409.00608", "identifiers": {"arxiv": "2409.00608"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.00608"}], "name": "TinyAgent: Function Calling at the Edge", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-25", "url": "https://arxiv.org/abs/2409.00608"}
{"aliases": [], "authors": ["Zhongze Tang", "Mengmei Ye", "Yao Liu", "Sheng Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27ac1bc0795cd8923a60de3cbe8aacf5f79a5fbeedfcfad6aae6f9457e16a9db", "submitted_at": "2024-09-03T08:47:17Z", "updated_at": "2024-09-03T08:47:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.01710v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2409.01710", "identifiers": {"arxiv": "2409.01710"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.01710"}], "name": "Privacy-Preserving Multimedia Mobile Cloud Computing Using Protective Perturbation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-03", "url": "https://arxiv.org/abs/2409.01710"}
{"aliases": [], "authors": ["Segev Shlomov", "Ben wiesel", "Aviad Sela", "Ido Levy", "Liane Galanti", "Roy Abitbol"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cdb228e652218d7f8981d9d16ca29b465004675c510d1e9c5c7b543bec9dc454", "submitted_at": "2024-09-03T14:17:09Z", "updated_at": "2024-09-03T14:17:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.01927v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2409.01927", "identifiers": {"arxiv": "2409.01927"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.01927"}], "name": "From Grounding to Planning: Benchmarking Bottlenecks in Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates web-agent planning and grounding evaluations on Mind2Web to analyze which components limit task performance.", "topics": ["evaluation", "grounding", "planning"], "updated_at": "2024-09-03", "url": "https://arxiv.org/abs/2409.01927"}
{"aliases": [], "authors": ["John Li", "Shehab Sarar Ahmed", "Deepak Nair"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7af0ea04235e191b09892aa37932bd4465d0141c27cbd14bde671b4726f2927", "submitted_at": "2024-09-04T05:19:57Z", "updated_at": "2024-09-10T08:20:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.02453v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.02453", "identifiers": {"arxiv": "2409.02453"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.02453"}], "name": "FrameCorr: Adaptive, Autoencoder-based Neural Compression for Video Reconstruction in Resource and Timing Constrained Network Settings", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-10", "url": "https://arxiv.org/abs/2409.02453"}
{"aliases": [], "authors": ["Jun Ji", "Zichen Xi", "Bernadeta R. Srijanto", "Ivan I. Kravchenko", "Ming Jin", "Wenjie Xiong", "Linbo Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c26c4316ca4c7129407a1379ed660b2dbbc2f2caa4438ddb464be4f01643801", "submitted_at": "2024-09-04T13:21:43Z", "updated_at": "2024-09-04T13:21:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.02689v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2409.02689", "identifiers": {"arxiv": "2409.02689"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.02689"}], "name": "Frequency-domain Parallel Computing Using Single On-Chip Nonlinear Acoustic-wave Device", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-04", "url": "https://arxiv.org/abs/2409.02689"}
{"aliases": [], "authors": ["Jingwei Zhang", "Thomas Lampe", "Abbas Abdolmaleki", "Jost Tobias Springenberg", "Martin Riedmiller"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49e23d0a100aae6e04b22a0bfea6856a7f802e8d2209e15c5ecb7dd340901505", "submitted_at": "2024-09-05T10:38:16Z", "updated_at": "2024-09-05T10:38:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.03402v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2409.03402", "identifiers": {"arxiv": "2409.03402"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.03402"}], "name": "Game On: Towards Language Models as RL Experimenters", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-05", "url": "https://arxiv.org/abs/2409.03402"}
{"aliases": [], "authors": ["Pawel Horodecki", "Łukasz Rudnicki", "Karol Życzkowski"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c6a9797a9dfa2e2cf87a1b25dd2a62b7a4765039370284f3b207c6e3000c6ed4", "submitted_at": "2024-09-06T19:05:15Z", "updated_at": "2024-09-06T19:05:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.04566v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.04566", "identifiers": {"arxiv": "2409.04566"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.04566"}], "name": "Multipartite entanglement", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-06", "url": "https://arxiv.org/abs/2409.04566"}
{"aliases": [], "authors": ["Ruilin Yao", "Shengwu Xiong", "Yichen Zhao", "Yi Rong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd321f0122de3a236019ddd3d0cd399eaa8837040c477f3fb8a8ccc6f00977ff", "submitted_at": "2024-09-08T07:08:58Z", "updated_at": "2024-09-08T07:08:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.04999v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2409.04999", "identifiers": {"arxiv": "2409.04999"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.04999"}], "name": "Visual Grounding with Multi-modal Conditional Adaptation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-08", "url": "https://arxiv.org/abs/2409.04999"}
{"aliases": [], "authors": ["Markus Knauer", "Alin Albu-Schäffer", "Freek Stulp", "João Silvério"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c9d7be59227ab5e9c4f98e80c87394229097a78b0b8f7a4292286c6b6ae5b51", "submitted_at": "2024-09-09T14:22:19Z", "updated_at": "2025-02-21T08:46:59Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.05655v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2409.05655", "identifiers": {"arxiv": "2409.05655"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.05655"}], "name": "Interactive incremental learning of generalizable skills with local trajectory modulation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-21", "url": "https://arxiv.org/abs/2409.05655"}
{"aliases": [], "authors": ["Zora Zhiruo Wang", "Jiayuan Mao", "Daniel Fried", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4c142be3a9c37bbb0d4f401a02de25bd25b4c3968f42ad9a61513abff2b5460", "submitted_at": "2024-09-11T17:21:00Z", "updated_at": "2024-09-11T17:21:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.07429v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.07429", "identifiers": {"arxiv": "2409.07429"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.07429"}], "name": "Agent Workflow Memory", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Agent Workflow Memory extracts reusable task routines and selectively supplies them to guide later web actions. It studies both offline workflow induction from training examples and online induction from test experience on Mind2Web and WebArena.", "topics": ["memory", "workflow", "web-navigation"], "updated_at": "2024-09-11", "url": "https://arxiv.org/abs/2409.07429"}
{"aliases": [], "authors": ["Liqiang Jing", "Zhehui Huang", "Xiaoyang Wang", "Wenlin Yao", "Wenhao Yu", "Kaixin Ma", "Hongming Zhang", "Xinya Du", "Dong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "96be578ba08ac6c865bdbcf2bd08abf44a81ae1df1fc222e9c5ab59ba5f79a22", "submitted_at": "2024-09-12T02:08:00Z", "updated_at": "2025-04-11T14:12:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.07703v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2409.07703", "identifiers": {"arxiv": "2409.07703"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.07703"}], "name": "DSBench: How Far Are Data Science Agents from Becoming Data Science Experts?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-11", "url": "https://arxiv.org/abs/2409.07703"}
{"aliases": [], "authors": ["Rogerio Bonatti", "Dan Zhao", "Francesco Bonacci", "Dillon Dupont", "Sara Abdali", "Yinheng Li", "Yadong Lu", "Justin Wagle", "Kazuhito Koishida", "Arthur Bucker", "Lawrence Jang", "Zack Hui"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe77f7f8ffd7490192ab54b7f6d49c0cb5b10be6df0a645bfb181b3be724118a", "submitted_at": "2024-09-12T17:56:43Z", "updated_at": "2024-09-13T20:17:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.08264v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.08264", "identifiers": {"arxiv": "2409.08264"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.08264"}], "name": "Windows Agent Arena: Evaluating Multi-Modal OS Agents at Scale", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-09-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adapts real-desktop evaluation to Windows applications and supports parallel execution of benchmark tasks.", "topics": ["evaluation", "cross-app"], "updated_at": "2024-09-13", "url": "https://arxiv.org/abs/2409.08264"}
{"aliases": [], "authors": ["Himanshu Gaurav Singh", "Antonio Loquercio", "Carmelo Sferrazza", "Jane Wu", "Haozhi Qi", "Pieter Abbeel", "Jitendra Malik"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a667d3e021d3b9e7154c329e6dfb72bc8d78b32ce6411c37d39bc8e7efae9134", "submitted_at": "2024-09-12T17:59:07Z", "updated_at": "2024-09-12T17:59:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.08273v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2409.08273", "identifiers": {"arxiv": "2409.08273"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.08273"}], "name": "Hand-Object Interaction Pretraining from Videos", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-12", "url": "https://arxiv.org/abs/2409.08273"}
{"aliases": [], "authors": ["Andreas Plesner", "Tobias Vontobel", "Roger Wattenhofer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97b1d955849b09ccb8a8021e909d307852b5de189623555a3d0a0e8d32aacad9", "submitted_at": "2024-09-13T13:47:12Z", "updated_at": "2024-09-13T13:47:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.08831v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2409.08831", "identifiers": {"arxiv": "2409.08831"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.08831"}], "name": "Breaking reCAPTCHAv2", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-13", "url": "https://arxiv.org/abs/2409.08831"}
{"aliases": [], "authors": ["Kelin Fu", "Yang Tian", "Kaigui Bian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b46dabe4135f0caca15d54f84cfe805cdb7fe7f1abcce6b004483d99eb21e637", "submitted_at": "2024-09-14T07:54:25Z", "updated_at": "2024-09-14T07:54:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.09354v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.09354", "identifiers": {"arxiv": "2409.09354"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.09354"}], "name": "PeriGuru: A Peripheral Robotic Mobile App Operation Assistant based on GUI Image Understanding and Prompting with LLM", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "PeriGuru interprets smartphone GUI screenshots with computer vision and uses a language model to decide the next action. A peripheral robotic arm executes the phone interactions, providing an alternative mechanism for operating mobile apps.", "topics": ["mobile-navigation", "screen-understanding", "automation"], "updated_at": "2024-09-14", "url": "https://arxiv.org/abs/2409.09354"}
{"aliases": [], "authors": ["Haichuan Hu", "Ye Shang", "Guolin Xu", "Congqing He", "Quanjun Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6806b863d40cb4bb79c45c87409a87a2def1e57b6eb6457a787ee2f645fd899a", "submitted_at": "2024-09-16T06:51:32Z", "updated_at": "2024-12-17T13:16:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.10033v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2409.10033", "identifiers": {"arxiv": "2409.10033"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.10033"}], "name": "Can GPT-O1 Kill All Bugs? An Evaluation of GPT-Family LLMs on QuixBugs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-17", "url": "https://arxiv.org/abs/2409.10033"}
{"aliases": [], "authors": ["Mobina Shahbandeh", "Parsa Alian", "Noor Nashid", "Ali Mesbah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5bdbd80e252cfa37dc62d666a202e647f93813549c55dde82daf36afe309c7f2", "submitted_at": "2024-09-16T21:18:39Z", "updated_at": "2024-09-16T21:18:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.10741v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.10741", "identifiers": {"arxiv": "2409.10741"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.10741"}], "name": "NaviQAte: Functionality-Guided Web Application Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Frames web-application exploration as functionality-oriented question answering, generating navigation actions from multimodal context.", "topics": ["web-navigation", "automation"], "updated_at": "2024-09-16", "url": "https://arxiv.org/abs/2409.10741"}
{"aliases": [], "authors": ["Shuqing Li", "Binchang Li", "Yepang Liu", "Cuiyun Gao", "Jianping Zhang", "Shing-Chi Cheung", "Michael R. Lyu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a32462e4cb09763750f8f22d7f4be32a6bdfe3db6c9b3de9c386dc7cd6b6f30", "submitted_at": "2024-09-17T00:58:00Z", "updated_at": "2025-10-01T17:58:44Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2409.10811v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2409.10811", "identifiers": {"arxiv": "2409.10811"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.10811"}], "name": "Grounded GUI Understanding for Vision-Based Spatial Intelligent Agent: Exemplified by Extended Reality Apps", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced. The described environment is XR/VR applications; the operating system is not specified in the abstract.", "published_at": "2024-09-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Orienter detects interactable GUI elements in XR applications by interpreting scene context and revisiting candidate elements through a validation and reflection loop. It addresses open-vocabulary element types and context-dependent interactability.", "topics": ["screen-understanding", "grounding"], "updated_at": "2025-10-01", "url": "https://arxiv.org/abs/2409.10811"}
{"aliases": [], "authors": ["Zeyi Liao", "Lingbo Mo", "Chejian Xu", "Mintong Kang", "Jiawei Zhang", "Chaowei Xiao", "Yuan Tian", "Bo Li", "Huan Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "897cdbb6aa52478b260ecba830e0a13527c2614454f7c766c9f5c8c9bdbfd4e3", "submitted_at": "2024-09-17T15:49:44Z", "updated_at": "2025-03-12T20:54:00Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2409.11295v5"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2409.11295", "identifiers": {"arxiv": "2409.11295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.11295"}], "name": "EIA: Environmental Injection Attack on Generalist Web Agents for Privacy Leakage", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-09-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates privacy-directed webpage instruction injection against generalist web agents using PII-related interaction steps from Mind2Web, including attempts to expose specific user information or entire requests.", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2025-03-12", "url": "https://arxiv.org/abs/2409.11295"}
{"aliases": [], "authors": ["Wayne Chi", "Ameet Talwalkar", "Chris Donahue"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2f59976cec7adb9f9c88bd79bdfbeed040e4874b6d9ab0c229136f98101aa98", "submitted_at": "2024-09-18T16:04:10Z", "updated_at": "2024-10-06T21:25:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.12089v3"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2409.12089", "identifiers": {"arxiv": "2409.12089"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.12089"}], "name": "The Impact of Element Ordering on LM Agent Performance", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2024-09-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies how the ordering of UI elements in an agent observation affects web and desktop performance. A pixel-based element detector and dimensionality-reduction ordering are evaluated on OmniACT when document hierarchy is unavailable.", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": "2024-10-06", "url": "https://arxiv.org/abs/2409.12089"}
{"aliases": [], "authors": ["Maria Wang", "Srinivas Sunkara", "Gilles Baechler", "Jason Lin", "Yun Zhu", "Fedir Zubach", "Lei Shu", "Jindong Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ca44e6c98c09a2261b2cf8f3b6d18584b2d2471934b2dcac8d5b72c17b32e61", "submitted_at": "2024-09-06T18:44:25Z", "updated_at": "2024-09-24T18:38:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.13711v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.13711", "identifiers": {"arxiv": "2409.13711"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.13711"}], "name": "WebQuest: A Benchmark for Multimodal QA on Web Page Sequences", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces question answering over individual screens, multiple web pages, and navigation traces to test information synthesis across views.", "topics": ["evaluation", "screen-understanding"], "updated_at": "2024-09-24", "url": "https://arxiv.org/abs/2409.13711"}
{"aliases": [], "authors": ["Longxi Gao", "Li Zhang", "Shihe Wang", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Shangguang Wang", "Yuanchun Li", "Mengwei Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4bb00ad1be49c9bd51881827556b3a168d34c69a45e31053406f780ad2a62af1", "submitted_at": "2024-09-22T06:45:38Z", "updated_at": "2025-11-25T02:43:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.14337v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.14337", "identifiers": {"arxiv": "2409.14337"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.14337"}], "name": "MobileViews: A Million-scale and Diverse Mobile GUI Dataset", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-09-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds mobile screenshot and view-hierarchy pairs through automated application traversal and evaluates their usefulness for GUI grounding.", "topics": ["screen-understanding", "data-generation"], "updated_at": "2025-11-25", "url": "https://arxiv.org/abs/2409.14337"}
{"aliases": [], "authors": ["Qinzhuo Wu", "Weikai Xu", "Wei Liu", "Tao Tan", "Jianfeng Liu", "Ang Li", "Jian Luan", "Bin Wang", "Shuo Shang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc22337ef29946dffcf7149df1008fd6cf02350c68d21ec0386e6485f5a14ebf", "submitted_at": "2024-09-23T08:47:54Z", "updated_at": "2024-10-03T05:23:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.14818v2"}, "publication": {"authors": ["Qinzhuo Wu", "Weikai Xu", "Wei Liu", "Tao Tan", "Jianfeng Liu", "Ang Li", "Jian Luan", "Bin Wang", "Shuo Shang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.599", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.599/", "https://aclanthology.org/2024.findings-emnlp.599.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the publisher PDF, page 1. The landing page and its BibTeX list the fifth author as Liujian Liujianfeng; the PDF names Jianfeng Liu.", "pages": "10231-10251", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.599.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "MobileVLM: A Vision-Language Model for Better Intra- and Inter-UI Understanding", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.599/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:52Z", "role": "primary", "sha256": "c7f0e00d6102bac259b21a9422fba11f3da72da06535f37581b2670082c82a7b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-emnlp.599/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:02.000149+00:00", "role": "primary", "sha256": "131e26bc7168b529a0db7c4bda0747cdf318b7ef16bfaf3b4dede3826cb85c1a", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2024.findings-emnlp.599.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.14818", "identifiers": {"arxiv": "2409.14818"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.14818"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-emnlp.599/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-emnlp.599.pdf"}], "name": "MobileVLM: A Vision-Language Model for Better Intra- and Inter-UI Understanding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileVLM adds two pretraining stages and four UI tasks to learn element detail and relationships between successive screens. Its Mobile3M dataset contains three million Chinese mobile UI pages connected by recorded transition actions.", "topics": ["screen-understanding", "data-generation", "action-prediction"], "updated_at": "2024-10-03", "url": "https://arxiv.org/abs/2409.14818"}
{"aliases": [], "authors": ["Brian Tang", "Kang G. Shin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "377d46c738f1a6f3e1c22ea99971a2fcfcf25ded82b2d2a24f728e614e22a1b3", "submitted_at": "2024-09-23T18:06:32Z", "updated_at": "2024-09-23T18:06:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.15441v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.15441", "identifiers": {"arxiv": "2409.15441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.15441"}], "name": "Steward: Natural Language Web Automation", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-09-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Steward, a natural-language browser automation system that repeatedly plans and executes website actions. The paper examines completion detection, state representation, responsiveness, and caching in real-site tasks.", "topics": ["web-navigation", "automation", "efficiency"], "updated_at": "2024-09-23", "url": "https://arxiv.org/abs/2409.15441"}
{"aliases": [], "authors": ["Tianyue Ou", "Frank F. Xu", "Aman Madaan", "Jiarui Liu", "Robert Lo", "Abishek Sridhar", "Sudipta Sengupta", "Dan Roth", "Graham Neubig", "Shuyan Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6c19b869b10f0b4c985b88b032f78e0e72f83e7ca8d9f346417054cd2a9002e3", "submitted_at": "2024-09-24T00:51:45Z", "updated_at": "2024-11-27T16:34:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.15637v2"}, "publication": {"authors": ["Tianyue Ou", "Frank F. Xu", "Aman Madaan", "Jiarui Liu", "Robert Lo", "Abishek Sridhar", "Sudipta Sengupta", "Dan Roth", "Graham Neubig", "Shuyan Zhou"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-2908", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "91618-91652", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/a6a6891cf1dfc64d664f086cf5976e93-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Synatra: Turning Indirect Knowledge into Direct Demonstrations for Digital Agents at Scale", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:04Z", "role": "primary", "sha256": "bf42d07f0959d7ad5ed0f4e00c34ef6cb7e8b1f10f17ac1e7a0ee1951523f821", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.15637", "identifiers": {"arxiv": "2409.15637"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.15637"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/a6a6891cf1dfc64d664f086cf5976e93-Paper-Conference.pdf"}], "name": "Synatra: Turning Indirect Knowledge into Direct Demonstrations for Digital Agents at Scale", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Synatra converts indirect procedural knowledge, including human-oriented tutorials, into direct demonstrations for digital agents. It trains a 7B CodeLlama on 100,000 synthesized demonstrations and evaluates the resulting policy on Mind2Web, MiniWoB++, and WebArena.", "topics": ["data-generation", "imitation-learning", "web-navigation"], "updated_at": "2024-11-27", "url": "https://arxiv.org/abs/2409.15637"}
{"aliases": [], "authors": ["Xuhui Zhou", "Hyunwoo Kim", "Faeze Brahman", "Liwei Jiang", "Hao Zhu", "Ximing Lu", "Frank Xu", "Bill Yuchen Lin", "Yejin Choi", "Niloofar Mireshghallah", "Ronan Le Bras", "Maarten Sap"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64913bd951fa6782e622c567c82b9f503f55ec5ef18355f4824c84d5448847e0", "submitted_at": "2024-09-24T19:47:21Z", "updated_at": "2025-08-30T18:01:52Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2409.16427v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2409.16427", "identifiers": {"arxiv": "2409.16427"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.16427"}], "name": "HAICOSYSTEM: An Ecosystem for Sandboxing Safety Risks in Human-AI Interactions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-30", "url": "https://arxiv.org/abs/2409.16427"}
{"aliases": [], "authors": ["Junting Lu", "Zhiyang Zhang", "Fangkai Yang", "Jue Zhang", "Lu Wang", "Chao Du", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "618f41c2253a532780766b38396f241a48b587c580d6b4d7065d1519e6149751", "submitted_at": "2024-09-25T17:58:08Z", "updated_at": "2025-05-19T16:12:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.17140v2"}, "publication": {"authors": ["Junting Lu", "Zhiyang Zhang", "Fangkai Yang", "Jue Zhang", "Lu Wang", "Chao Du", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.381", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.381/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "7711-7743", "pdf_url": "https://aclanthology.org/2025.acl-long.381.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AXIS: Efficient Human-Agent-Computer Interaction with API-First LLM-Based Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.381/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:52Z", "role": "primary", "sha256": "31ab26306d5c41b25731db1ad7f73495330e637c1694c9da644f4e238a73917b", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.381/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.17140", "identifiers": {"arxiv": "2409.17140"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.17140"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.381/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.381.pdf"}], "name": "AXIS: Efficient Human-Agent-Computer Interaction with API-First LLM-Based Agents", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-09-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AXIS, an API-first application agent that can use UI actions and expand its available APIs through automated application exploration. The work evaluates the approach on Microsoft Word tasks.", "topics": ["tool-integration", "automation", "efficiency"], "updated_at": "2025-05-19", "url": "https://arxiv.org/abs/2409.17140"}
{"aliases": [], "authors": ["Fan-Yun Sun", "S. I. Harini", "Angela Yi", "Yihan Zhou", "Alex Zook", "Jonathan Tremblay", "Logan Cross", "Jiajun Wu", "Nick Haber"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f85da331d3960044260b572a9a00b8b1bf5a5bfa6e02079817b824c8b2d1d5e1", "submitted_at": "2024-09-26T09:00:30Z", "updated_at": "2024-11-11T08:16:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.17652v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2409.17652", "identifiers": {"arxiv": "2409.17652"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.17652"}], "name": "FactorSim: Generative Simulation via Factorized Representation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-11", "url": "https://arxiv.org/abs/2409.17652"}
{"aliases": [], "authors": ["Sabrina Haque", "Christoph Csallner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d927b7644f0c6b3a49fbdf1aa976a00f2409adf9a72e1c04333438f66f1014b", "submitted_at": "2024-09-26T17:01:33Z", "updated_at": "2024-10-07T19:04:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.18060v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2409.18060", "identifiers": {"arxiv": "2409.18060"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.18060"}], "name": "Inferring Alt-text For UI Icons With Large Language Models During App Development", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. This is a UI accessibility and element-description method; autonomous GUI control is not evaluated.", "published_at": "2024-09-26", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces IconDesc, which generates mobile icon alt-text from partial UI context such as element identifiers, bounds, nearby text and hierarchy. Model evaluation and a user study examine the resulting accessibility descriptions.", "topics": ["screen-understanding", "ui-design", "evaluation"], "updated_at": "2024-10-07", "url": "https://arxiv.org/abs/2409.18060"}
{"aliases": [], "authors": ["Hyungjoo Chae", "Taeyoon Kwon", "Seungjun Moon", "Yongho Song", "Dongjin Kang", "Kai Tzu-iunn Ong", "Beong-woo Kwak", "Seonghyeon Bae", "Seung-won Hwang", "Jinyoung Yeo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "128dd462c3d59ea76f770ccd855dac55d0adb1ae58df772852ccc665913b2aff", "submitted_at": "2024-09-29T14:14:25Z", "updated_at": "2024-10-04T08:48:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.19715v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2409.19715", "identifiers": {"arxiv": "2409.19715"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.19715"}], "name": "Coffee-Gym: An Environment for Evaluating and Improving Natural Language Feedback on Erroneous Code", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-04", "url": "https://arxiv.org/abs/2409.19715"}
{"aliases": [], "authors": ["Haotian Zhang", "Mingfei Gao", "Zhe Gan", "Philipp Dufter", "Nina Wenzel", "Forrest Huang", "Dhruti Shah", "Xianzhi Du", "Bowen Zhang", "Yanghao Li", "Sam Dodge", "Keen You", "Zhen Yang", "Aleksei Timofeev", "Mingze Xu", "Hong-You Chen", "Jean-Philippe Fauconnier", "Zhengfeng Lai", "Haoxuan You", "Zirui Wang", "Afshin Dehghan", "Peter Grasch", "Yinfei Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2e066b476434bf2899de8dc6b96d6e10219e731f47d848255702d68c99bf5d3e", "submitted_at": "2024-09-30T17:59:34Z", "updated_at": "2024-09-30T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.20566v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "arxiv:2409.20566", "identifiers": {"arxiv": "2409.20566"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2409.20566"}], "name": "MM1.5: Methods, Analysis & Insights from Multimodal LLM Fine-tuning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-09-30", "url": "https://arxiv.org/abs/2409.20566"}
{"aliases": [], "authors": ["Wenyue Hua", "Mengting Wan", "Shashank Vadrevu", "Ryan Nadel", "Yongfeng Zhang", "Chi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f4c873d1dd443ffc18fc4b63595fdfc726eb84388eddbfced0dcde67c5e19a7", "submitted_at": "2024-09-30T16:52:51Z", "updated_at": "2024-09-30T16:52:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.00079v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2410.00079", "identifiers": {"arxiv": "2410.00079"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.00079"}], "name": "Interactive Speculative Planning: Enhance Agent Efficiency through Co-design of System and User Interface", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-30", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Studies agent-system and interface co-design that incorporates user interruptions and intermediate feedback to reduce planning latency.", "topics": ["planning", "efficiency", "human-in-the-loop"], "updated_at": "2024-09-30", "url": "https://arxiv.org/abs/2410.00079"}
{"aliases": [], "authors": ["Yi-Hao Peng", "Faria Huq", "Yue Jiang", "Jason Wu", "Amanda Xin Yue Li", "Jeffrey Bigham", "Amy Pavel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a85dca895f00b442a60123c7c85cd26a90e3d1c2a810c91c32385f877bcd755a", "submitted_at": "2024-09-30T19:55:54Z", "updated_at": "2024-09-30T19:55:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.00201v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2410.00201", "identifiers": {"arxiv": "2410.00201"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.00201"}], "name": "DreamStruct: Understanding Slides and User Interfaces via Synthetic Data Generation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Generates labeled slides and interface visuals through code to support visual element detection, description, and classification with less manual annotation.", "topics": ["data-generation", "screen-understanding"], "updated_at": "2024-09-30", "url": "https://arxiv.org/abs/2410.00201"}
{"aliases": [], "authors": ["Dawit Zeleke Admassu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8c65f99aa129a9aa4b151238575e34b771172e633baa032abc2412808e3ad042", "submitted_at": "2024-10-01T04:35:44Z", "updated_at": "2024-10-01T04:35:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.00395v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.00395", "identifiers": {"arxiv": "2410.00395"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.00395"}], "name": "Performance Improvement of IaaS Type of Cloud Computing Using Virtualization Technique", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-01", "url": "https://arxiv.org/abs/2410.00395"}
{"aliases": [], "authors": ["Shaoqing Zhang", "Zhuosheng Zhang", "Kehai Chen", "Xinbei Ma", "Muyun Yang", "Tiejun Zhao", "Min Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f370d105dc1fc1a377fd1b1fd5d1fc3d0f035694cc7c364e6326c587607c98c6", "submitted_at": "2024-10-01T07:49:24Z", "updated_at": "2024-12-19T14:29:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.00467v3"}, "publication": {"authors": ["Shaoqing Zhang", "Zhuosheng Zhang", "Kehai Chen (陈科海)", "Xinbei Ma", "Muyun Yang (杨沐昀)", "Tiejun Zhao (赵铁军)", "Min Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.70", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.70/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "1304-1320", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.70.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Dynamic Planning for LLM-based Graphical User Interface Automation", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.70/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:53Z", "role": "primary", "sha256": "1d89601c2a7a81756b9e7441c48a312132f385e5370780507d27c9848aa1176d", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-emnlp.70/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.00467", "identifiers": {"arxiv": "2410.00467"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.00467"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-emnlp.70/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-emnlp.70.pdf"}], "name": "Dynamic Planning for LLM-based Graphical User Interface Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Dynamic Planning of Thoughts revises a mobile GUI plan using current environmental feedback and execution history. The study compares this planning strategy with long-dialogue prompting and examines its behavior on unfamiliar tasks and different language-model backbones.", "topics": ["planning", "interaction-history", "mobile-navigation", "robustness"], "updated_at": "2024-12-19", "url": "https://arxiv.org/abs/2410.00467"}
{"aliases": [], "authors": ["Ruhana Azam", "Tamer Abuelsaad", "Aditya Vempaty", "Ashish Jagmohan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "69febc7aabfe38a6da439f1b4f673e0c36bc34c6afd1e8b06481433c1f43b1aa", "submitted_at": "2024-10-01T13:43:55Z", "updated_at": "2024-10-11T15:42:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.00689v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.00689", "identifiers": {"arxiv": "2410.00689"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.00689"}], "name": "Multimodal Auto Validation For Self-Refinement in Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies text and visual validation for web agents and uses the resulting validator to detect failures and guide workflow self-refinement.", "topics": ["evaluation", "robustness", "web-navigation"], "updated_at": "2024-10-11", "url": "https://arxiv.org/abs/2410.00689"}
{"aliases": [], "authors": ["Tomas Goldsack", "Yang Wang", "Chenghua Lin", "Chung-Chi Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9fffc07c46b29aa045007dc43a4c5dbfbd7cbb42a645f5530795d7ff785d439b", "submitted_at": "2024-10-01T20:03:22Z", "updated_at": "2024-10-01T20:03:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.01039v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2410.01039", "identifiers": {"arxiv": "2410.01039"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.01039"}], "name": "From Facts to Insights: A Study on the Generation and Evaluation of Analytical Reports for Deciphering Earnings Calls", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-01", "url": "https://arxiv.org/abs/2410.01039"}
{"aliases": [], "authors": ["Maya Pavlova", "Erik Brinkman", "Krithika Iyer", "Vitor Albiero", "Joanna Bitton", "Hailey Nguyen", "Joe Li", "Cristian Canton Ferrer", "Ivan Evtimov", "Aaron Grattafiori"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5adda49ab71b984a273afbe44156ffb34c2d0d617a0e49e259149062ffa38a8c", "submitted_at": "2024-10-02T14:47:05Z", "updated_at": "2024-10-02T14:47:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.01606v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2410.01606", "identifiers": {"arxiv": "2410.01606"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.01606"}], "name": "Automated Red Teaming with GOAT: the Generative Offensive Agent Tester", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-02", "url": "https://arxiv.org/abs/2410.01606"}
{"aliases": [], "authors": ["Nikolai Rozanov", "Marek Rei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a2e16e978f50c8d7161a11a11fba6aeda9617663de1746422151d38c4874cc9", "submitted_at": "2024-09-21T05:54:35Z", "updated_at": "2025-04-08T06:37:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.02810v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.02810", "identifiers": {"arxiv": "2410.02810"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.02810"}], "name": "StateAct: Enhancing LLM Base Agents via Self-prompting and State-tracking", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-09-21", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Maintains explicit task goals and evolving state descriptions to improve base-agent decisions across interactive tasks, including web shopping.", "topics": ["planning", "memory"], "updated_at": "2025-04-08", "url": "https://arxiv.org/abs/2410.02810"}
{"aliases": [], "authors": ["Shikhar Murty", "Hao Zhu", "Dzmitry Bahdanau", "Christopher D. Manning"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e227f3686b3234745ca1749cb8fb7c2b19383ac9dfebf68976f61f07d1ddf3b", "submitted_at": "2024-10-03T18:56:51Z", "updated_at": "2025-02-05T18:56:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.02907v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.02907", "identifiers": {"arxiv": "2410.02907"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.02907"}], "name": "NNetNav: Unsupervised Learning of Browser Agents Through Environment Interaction in the Wild", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Generates browser-agent demonstrations through autonomous exploration and retrospective task labeling, pruning trajectories that lack meaningful subtasks.", "topics": ["data-generation", "web-navigation"], "updated_at": "2025-02-05", "url": "https://arxiv.org/abs/2410.02907"}
{"aliases": [], "authors": ["Boyu Gou", "Ruohan Wang", "Boyuan Zheng", "Yanan Xie", "Cheng Chang", "Yiheng Shu", "Huan Sun", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08ece826d135b9ab9dfed2a3baf9deeab99489120da749cd67517b636649c5a6", "submitted_at": "2024-10-07T17:47:50Z", "updated_at": "2025-06-17T15:06:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.05243v3"}, "publication": {"authors": ["Boyu Gou", "Demi Ruohan Wang", "Boyuan Zheng", "Yanan Xie", "Cheng Chang", "Yiheng Shu", "Huan Sun", "Yu Su"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "30851-30883", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/4ca0e369689dadb25a5345ba9755ad6f-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:01Z", "role": "primary", "sha256": "4e7384659700d8bc8e1f56de671c0d943f15b8c682bdff0fbfb20f41c707b7ff", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.05243", "identifiers": {"arxiv": "2410.05243"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.05243"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/4ca0e369689dadb25a5345ba9755ad6f-Paper-Conference.pdf"}], "name": "Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces UGround, a visual target-localization model trained with large-scale synthetic references to elements in GUI screenshots.", "topics": ["grounding", "data-generation"], "updated_at": "2025-06-17", "url": "https://arxiv.org/abs/2410.05243"}
{"aliases": [], "authors": ["Kamel Alrashedy", "Pradyumna Tambwekar", "Zulfiqar Zaidi", "Megan Langwasser", "Wei Xu", "Matthew Gombolay"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4af292de2d267cf8275427d89059320ef98def65ade8ebaedec76d1c985de9a7", "submitted_at": "2024-10-07T02:44:50Z", "updated_at": "2025-02-28T04:28:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.05340v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2410.05340", "identifiers": {"arxiv": "2410.05340"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.05340"}], "name": "Generating CAD Code with Vision-Language Models for 3D Designs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-28", "url": "https://arxiv.org/abs/2410.05340"}
{"aliases": [], "authors": ["Martin Klissarov", "Devon Hjelm", "Alexander Toshev", "Bogdan Mazoure"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b8391e2d9593f8ab069c0422fe10268c149dc46e0be5ba7c80103042dbc64ee7", "submitted_at": "2024-10-08T03:12:57Z", "updated_at": "2024-10-08T03:12:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.05656v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}], "id": "arxiv:2410.05656", "identifiers": {"arxiv": "2410.05656"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.05656"}], "name": "On the Modeling Capabilities of Large Language Models for Sequential Decision Making", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-08", "url": "https://arxiv.org/abs/2410.05656"}
{"aliases": [], "authors": ["Adil Zouitine", "Matthieu Geist", "Emmanuel Rachelson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7611026e82d20ef42e4bca569b98c3d692caa795e146aa08037b91a18140f74c", "submitted_at": "2024-10-08T17:16:24Z", "updated_at": "2024-10-08T17:16:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.06212v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.06212", "identifiers": {"arxiv": "2410.06212"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.06212"}], "name": "Solving robust MDPs as a sequence of static RL problems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-08", "url": "https://arxiv.org/abs/2410.06212"}
{"aliases": [], "authors": ["Zaid Khan", "Elias Stengel-Eskin", "Jaemin Cho", "Mohit Bansal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c6e362eba1aa0f5a7f4b5496594ea7ac85d636441220734a68fe82d48de82af", "submitted_at": "2024-10-08T17:20:37Z", "updated_at": "2025-03-13T17:30:48Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.06215v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.06215", "identifiers": {"arxiv": "2410.06215"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.06215"}], "name": "DataEnvGym: Data Generation Agents in Teacher Environments with Student Feedback", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-13", "url": "https://arxiv.org/abs/2410.06215"}
{"aliases": [], "authors": ["Ido Levy", "Ben Wiesel", "Sami Marreed", "Alon Oved", "Avi Yaeli", "Nir Mashkif", "Segev Shlomov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ba2b599c2b3f09974ee41d329545b69773a3b408b6db0124c5b1073d5fe8f1cf", "submitted_at": "2024-10-09T09:13:38Z", "updated_at": "2026-06-04T09:59:28Z", "version": 7, "versioned_url": "https://arxiv.org/abs/2410.06703v7"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.06703", "identifiers": {"arxiv": "2410.06703"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.06703"}], "name": "ST-WebAgentBench: A Benchmark for Evaluating Safety and Trustworthiness in Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates web-task completion under explicit enterprise policies, measuring policy-compliant completion and safety violations separately.", "topics": ["safety-security", "evaluation"], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2410.06703"}
{"aliases": [], "authors": ["Manling Li", "Shiyu Zhao", "Qineng Wang", "Kangrui Wang", "Yu Zhou", "Sanjana Srivastava", "Cem Gokmen", "Tony Lee", "Li Erran Li", "Ruohan Zhang", "Weiyu Liu", "Percy Liang", "Li Fei-Fei", "Jiayuan Mao", "Jiajun Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9d700e93c9ee77106cf743902de68b20fffcbee4428bfcac35b114ac0971d6b4", "submitted_at": "2024-10-09T17:59:00Z", "updated_at": "2025-01-19T19:29:50Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.07166v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2410.07166", "identifiers": {"arxiv": "2410.07166"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.07166"}], "name": "Embodied Agent Interface: Benchmarking LLMs for Embodied Decision Making", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-19", "url": "https://arxiv.org/abs/2410.07166"}
{"aliases": [], "authors": ["Donghyun Lee", "Mo Tiwari"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f2386773b2ca1d721d7b412d84898f7e37e3bb0c924d7ab237e37b2b5701562f", "submitted_at": "2024-10-09T11:01:29Z", "updated_at": "2024-10-09T11:01:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.07283v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.07283", "identifiers": {"arxiv": "2410.07283"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.07283"}], "name": "Prompt Infection: LLM-to-LLM Prompt Injection within Multi-Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-09", "url": "https://arxiv.org/abs/2410.07283"}
{"aliases": [], "authors": ["Yiming Huang", "Jianwen Luo", "Yan Yu", "Yitong Zhang", "Fangyu Lei", "Yifan Wei", "Shizhu He", "Lifu Huang", "Xiao Liu", "Jun Zhao", "Kang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "83991dcc79a5efd3b401e66d81f5b7c385294ab108431e329e4d980ec5bd77e5", "submitted_at": "2024-10-09T18:00:05Z", "updated_at": "2024-10-11T00:53:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.07331v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.07331", "identifiers": {"arxiv": "2410.07331"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.07331"}], "name": "DA-Code: Agent Data Science Code Generation Benchmark for Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-11", "url": "https://arxiv.org/abs/2410.07331"}
{"aliases": [], "authors": ["Shion Mifune", "Taro Kanao", "Tetsufumi Tanamoto"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b05f14bc0bdb4438929b3afbc2d699f3e606ea71ebb28af5c1da5967bb4f75fd", "submitted_at": "2024-10-09T23:23:50Z", "updated_at": "2025-04-04T01:11:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.07481v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.07481", "identifiers": {"arxiv": "2410.07481"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.07481"}], "name": "Effects of dissipation in reservoir computing using a spin qubit array", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-04", "url": "https://arxiv.org/abs/2410.07481"}
{"aliases": [], "authors": ["Shuofei Qiao", "Runnan Fang", "Zhisong Qiu", "Xiaobin Wang", "Ningyu Zhang", "Yong Jiang", "Pengjun Xie", "Fei Huang", "Huajun Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "84a11f9b0e9188636595df693bc55c68e89e87388e7cbcd2720ba547e53a8371", "submitted_at": "2024-10-10T12:41:19Z", "updated_at": "2025-02-23T15:16:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.07869v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.07869", "identifiers": {"arxiv": "2410.07869"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.07869"}], "name": "Benchmarking Agentic Workflow Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-23", "url": "https://arxiv.org/abs/2410.07869"}
{"aliases": [], "authors": ["Saaket Agashe", "Jiuzhou Han", "Shuyu Gan", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "284888d16b049767f76448f23f5d35fda25b109fd67f450105950173ae81c97b", "submitted_at": "2024-10-10T17:43:51Z", "updated_at": "2024-10-10T17:43:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.08164v1"}, "publication": {"authors": ["Saaket Agashe", "Jiuzhou Han", "Shuyu Gan", "Jiachen Yang", "Ang Li", "Xin Wang"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "22924-22946", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/394c7c30ea87b5c3521b4d9e9d419071-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Agent S: An Open Agentic Framework that Uses Computers Like a Human", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:01Z", "role": "primary", "sha256": "032cfd2f08acc7f85921748e0546bf6e16d038da412c207105ca9f6d433c70b3", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.08164", "identifiers": {"arxiv": "2410.08164"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.08164"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/394c7c30ea87b5c3521b4d9e9d419071-Paper-Conference.pdf"}], "name": "Agent S: An Open Agentic Framework that Uses Computers Like a Human", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2024-10-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines hierarchical planning, retrieved experience, and a computer interaction interface to execute multistep GUI tasks.", "topics": ["planning", "memory"], "updated_at": "2024-10-10", "url": "https://arxiv.org/abs/2410.08164"}
{"aliases": [], "authors": ["Shengcao Cao", "Liang-Yan Gui", "Yu-Xiong Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c65c701bef939dc02007755b60d4eac76d68f3d9f45c952fe33da282a0007218", "submitted_at": "2024-10-10T17:59:55Z", "updated_at": "2025-10-16T07:50:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.08209v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.08209", "identifiers": {"arxiv": "2410.08209"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.08209"}], "name": "Emergent Visual Grounding in Large Multimodal Models Without Grounding Supervision", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-16", "url": "https://arxiv.org/abs/2410.08209"}
{"aliases": [], "authors": ["Wilson Yan", "Volodymyr Mnih", "Aleksandra Faust", "Matei Zaharia", "Pieter Abbeel", "Hao Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c6def90f5cea391601541b74c2d0361540067851880a37cf5e9488b94b4a759", "submitted_at": "2024-10-10T20:54:15Z", "updated_at": "2025-02-02T19:28:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.08368v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.08368", "identifiers": {"arxiv": "2410.08368"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.08368"}], "name": "ElasticTok: Adaptive Tokenization for Image and Video", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-02", "url": "https://arxiv.org/abs/2410.08368"}
{"aliases": [], "authors": ["Zhuohao Jerry Zhang", "Eldon Schoop", "Jeffrey Nichols", "Anuj Mahajan", "Amanda Swearngin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff649a0792864bd40b7bf2f825f6a8a815eb60cc795ed4dc0c6b55a942477758", "submitted_at": "2024-10-11T17:24:00Z", "updated_at": "2025-03-22T18:01:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.09006v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.09006", "identifiers": {"arxiv": "2410.09006"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.09006"}], "name": "From Interaction to Impact: Towards Safer AI Agents Through Understanding and Evaluating Mobile UI Operation Impacts", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Develops a taxonomy of mobile-action consequences and annotated interaction data to evaluate whether models recognize impactful or risky UI operations.", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2025-03-22", "url": "https://arxiv.org/abs/2410.09006"}
{"aliases": [], "authors": ["Maksym Andriushchenko", "Alexandra Souly", "Mateusz Dziemian", "Derek Duenas", "Maxwell Lin", "Justin Wang", "Dan Hendrycks", "Andy Zou", "Zico Kolter", "Matt Fredrikson", "Eric Winsor", "Jerome Wynne", "Yarin Gal", "Xander Davies"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "516386ba13b255df5469a4082d9361f76b32858e35a2f9561611ecdf31ca5aff", "submitted_at": "2024-10-11T17:39:22Z", "updated_at": "2025-04-18T14:30:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.09024v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.09024", "identifiers": {"arxiv": "2410.09024"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.09024"}], "name": "AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-18", "url": "https://arxiv.org/abs/2410.09024"}
{"aliases": [], "authors": ["Jinjin Shen", "Yan Lin", "Yijin Zhang", "Weibin Zhang", "Feng Shu", "Jun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb3c4a5cbbfe3b48af743f891f58fa52d2929483cb0639829ec23a3e6a733883", "submitted_at": "2024-10-14T01:25:56Z", "updated_at": "2024-10-14T01:25:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.10071v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.10071", "identifiers": {"arxiv": "2410.10071"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.10071"}], "name": "Content Caching-Assisted Vehicular Edge Computing Using Multi-Agent Graph Attention Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-14", "url": "https://arxiv.org/abs/2410.10071"}
{"aliases": [], "authors": ["Rui Zhang", "Ziyao Zhang", "Fengliang Zhu", "Jiajie Zhou", "Anyi Rao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43620f85e02f6624f1e3461eae77e0f6ceea15e2ebacbe843d6371b683f7a140", "submitted_at": "2024-10-14T14:47:32Z", "updated_at": "2024-10-15T08:56:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.10570v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.10570", "identifiers": {"arxiv": "2410.10570"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.10570"}], "name": "Mindalogue: LLM-Powered Nonlinear Interaction for Effective Learning and Task Exploration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-15", "url": "https://arxiv.org/abs/2410.10570"}
{"aliases": [], "authors": ["Zeru Shi", "Kai Mei", "Mingyu Jin", "Yongye Su", "Chaoji Zuo", "Wenyue Hua", "Wujiang Xu", "Yujie Ren", "Zirui Liu", "Mengnan Du", "Dong Deng", "Yongfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa26cd90a9bcd0ef72e1721630f84b03cad381f5dff9c82eebf589be75d984f2", "submitted_at": "2024-09-23T08:39:16Z", "updated_at": "2025-03-19T03:17:47Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2410.11843v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.11843", "identifiers": {"arxiv": "2410.11843"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.11843"}], "name": "From Commands to Prompts: LLM-based Semantic File System for AIOS", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-09-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-19", "url": "https://arxiv.org/abs/2410.11843"}
{"aliases": [], "authors": ["Pawel Pawlowski", "Krystian Zawistowski", "Wojciech Lapacz", "Adam Wiacek", "Marcin Skorupa", "Sebastien Postansque", "Jakub Hoscilowicz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f1594a51709f2a1283d3ac14734f08fa7894b436b85bafbbbe9b2eaba332129", "submitted_at": "2024-10-09T12:06:43Z", "updated_at": "2025-05-21T06:52:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.11871v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.11871", "identifiers": {"arxiv": "2410.11871"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.11871"}], "name": "TinyClick: Single-Turn Agent for Empowering GUI Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adapts a compact Florence-2 model to map user instructions to screen coordinates using multitask training and augmented data.", "topics": ["grounding", "efficiency"], "updated_at": "2025-05-21", "url": "https://arxiv.org/abs/2410.11871"}
{"aliases": [], "authors": ["Jakub Hoscilowicz", "Bartosz Maj", "Bartosz Kozakiewicz", "Oleksii Tymoshchuk", "Artur Janicki"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e804944da205076ac6cf88c7966c471697e526f657bda339e09df240257f7217", "submitted_at": "2024-10-09T14:49:02Z", "updated_at": "2024-10-17T07:12:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.11872v2"}, "publication": {"authors": ["Jakub Hoscilowicz", "Bartosz Maj", "Bartosz Kozakiewicz", "Oleksii Tymoshchuk", "Artur Janicki"], "date_basis": "Explicit publisher citation_publication_date: 2025/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 8], "doi": "10.18653/v1/2025.sigdial-1.38", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.sigdial-1.38/", "https://aclanthology.org/2025.sigdial-1.38.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the five-author list on page 1 of the publisher PDF. Landing-page citation metadata lists only Jakub Hoscilowicz and Artur Janicki.", "pages": "471-476", "pdf_url": "https://aclanthology.org/2025.sigdial-1.38.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "ClickAgent: Enhancing UI Location Capabilities of Autonomous Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.sigdial-1.38/", "venue": "Proceedings of the 26th Annual Meeting of the Special Interest Group on Discourse and Dialogue", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:53Z", "role": "primary", "sha256": "7c710272d48754149cc5947f5382b8123eac5855db7655410259c4a3129b8b5b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.sigdial-1.38/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:09.676954+00:00", "role": "primary", "sha256": "c5b5cd07e3ca2f2911e183c4752e01c721971c47fd20c836605452f2b80e5260", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2025.sigdial-1.38.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.11872", "identifiers": {"arxiv": "2410.11872"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.11872"}, {"rel": "publication", "url": "https://aclanthology.org/2025.sigdial-1.38/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.sigdial-1.38.pdf"}], "name": "ClickAgent: Enhancing UI Location Capabilities of Autonomous Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates multimodal reasoning and action planning from a dedicated UI localization model, with evaluation on Android devices.", "topics": ["mobile-navigation", "grounding"], "updated_at": "2024-10-17", "url": "https://arxiv.org/abs/2410.11872"}
{"aliases": [], "authors": ["Loris Gaven", "Clement Romac", "Thomas Carta", "Sylvain Lamprier", "Olivier Sigaud", "Pierre-Yves Oudeyer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8d7d809488791ef6d68dd11b2435b0eeb604e695b02eb2802f147b1c4ba90038", "submitted_at": "2024-10-16T11:59:27Z", "updated_at": "2026-01-29T16:31:07Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.12481v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.12481", "identifiers": {"arxiv": "2410.12481"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.12481"}], "name": "SAC-GLAM: Improving Online RL for LLM agents with Soft Actor-Critic and Hindsight Relabeling", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-29", "url": "https://arxiv.org/abs/2410.12481"}
{"aliases": [], "authors": ["Dewei Zhou", "Ji Xie", "Zongxin Yang", "Yi Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c44e6ed329a203f2d8c9439af466fce6671faef47c9109ff724fe81f3bcef91f", "submitted_at": "2024-10-16T15:34:13Z", "updated_at": "2025-12-02T12:08:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.12669v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.12669", "identifiers": {"arxiv": "2410.12669"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.12669"}], "name": "3DIS: Depth-Driven Decoupled Instance Synthesis for Text-to-Image Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-02", "url": "https://arxiv.org/abs/2410.12669"}
{"aliases": [], "authors": ["Ali Al-Bayaty", "Marek Perkowski"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "427c6c6d0402ca03b8fc7cfbd856be6b5d9459d859aec93433e3635b06c75d93", "submitted_at": "2024-10-17T00:29:29Z", "updated_at": "2024-10-17T00:29:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.13104v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.13104", "identifiers": {"arxiv": "2410.13104"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.13104"}], "name": "Cost-Effective Realization of n-Bit Toffoli Gates for IBM Quantum Computers Using the Bloch Sphere Approach and IBM Native Gates", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-17", "url": "https://arxiv.org/abs/2410.13104"}
{"aliases": [], "authors": ["Hyungjoo Chae", "Namyoung Kim", "Kai Tzu-iunn Ong", "Minju Gwak", "Gwanwoo Song", "Jihoon Kim", "Sunghwan Kim", "Dongha Lee", "Jinyoung Yeo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cae86bff4412bbfa7b969ecd60fe7b695f4b7c5ea1be6a67e97919df762d1108", "submitted_at": "2024-10-17T05:37:00Z", "updated_at": "2025-03-29T08:59:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.13232v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.13232", "identifiers": {"arxiv": "2410.13232"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13232"}], "name": "Web Agents with World Models: Learning and Leveraging Environment Dynamics in Web Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses a learned web world model to predict important state changes and compare candidate actions before executing them.", "topics": ["world-models", "planning"], "updated_at": "2025-03-29", "url": "https://arxiv.org/abs/2410.13232"}
{"aliases": [], "authors": ["Zichen Zhu", "Hao Tang", "Yansi Li", "Dingye Liu", "Hongshen Xu", "Kunyao Lan", "Danyang Zhang", "Yixuan Jiang", "Hao Zhou", "Chenrun Wang", "Situo Zhang", "Liangtai Sun", "Yixiao Wang", "Yuheng Sun", "Lu Chen", "Kai Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0ef02f9c31bf2f095330cf018e7525aecc7644eb1552fb6b511ffced3889122", "submitted_at": "2024-10-17T16:53:50Z", "updated_at": "2025-05-13T06:25:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.13757v3"}, "publication": {"authors": ["Zichen Zhu", "Hao Tang", "Yansi Li", "Dingye Liu", "Hongshen Xu", "Kunyao Lan", "Danyang Zhang", "Yixuan Jiang", "Hao Zhou", "Chenrun Wang", "Situo Zhang", "Liangtai Sun", "Yixiao Wang", "Yuheng Sun", "Lu Chen", "Kai Yu"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-demo.43", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-demo.43/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "535-549", "pdf_url": "https://aclanthology.org/2025.naacl-demo.43.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "MobA: Multifaceted Memory-Enhanced Adaptive Planning for Efficient Mobile Task Automation", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-demo.43/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:53Z", "role": "primary", "sha256": "a873b34264be14f64599c0897219ff2796ab74d111df8a2fdc667fd22ad87028", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-demo.43/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.13757", "identifiers": {"arxiv": "2410.13757"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13757"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-demo.43/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-demo.43.pdf"}], "name": "MobA: Multifaceted Memory-Enhanced Adaptive Planning for Efficient Mobile Task Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines adaptive mobile-task planning, reflection-based recovery, and multiple memory components, evaluated with MobBench and AndroidArena.", "topics": ["memory", "planning", "mobile-navigation"], "updated_at": "2025-05-13", "url": "https://arxiv.org/abs/2410.13757"}
{"aliases": [], "authors": ["Junpeng Liu", "Tianyue Ou", "Yifan Song", "Yuxiao Qu", "Wai Lam", "Chenyan Xiong", "Wenhu Chen", "Graham Neubig", "Xiang Yue"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a3bcd26dff4f1d21a8a2772929c0c97cd1cc180a7606b0744e630dabae94609", "submitted_at": "2024-10-17T17:48:54Z", "updated_at": "2024-11-06T08:29:22Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.13824v3"}, "publication": {"authors": ["Junpeng Liu", "Tianyue Ou", "Yifan Song", "Yuxiao Qu", "Wai Lam", "Chenyan Xiong", "Wenhu Chen", "Graham Neubig", "Xiang Yue"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "17971-18005", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/2da53cd1abdae59150e35f4693834f32-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Harnessing Webpage UIs for Text-Rich Visual Understanding", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "899100e1d0cad8d949c9075f62ab3506995587b01aab3981ac4e26eaf6d88339", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.13824", "identifiers": {"arxiv": "2410.13824"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13824"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/2da53cd1abdae59150e35f4693834f32-Paper-Conference.pdf"}], "name": "Harnessing Webpage UIs for Text-Rich Visual Understanding", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-10-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds MultiUI by generating instructions from webpage accessibility trees and pairing them with screenshots for multimodal training. Evaluation examines web-interface understanding, including VisualWebBench and Mind2Web, and transfer to other text-rich visual tasks.", "topics": ["data-generation", "screen-understanding", "grounding"], "updated_at": "2024-11-06", "url": "https://arxiv.org/abs/2410.13824"}
{"aliases": [], "authors": ["Ke Yang", "Yao Liu", "Sapana Chaudhary", "Rasool Fakoor", "Pratik Chaudhari", "George Karypis", "Huzefa Rangwala"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "09268d51b3d810ffc0024ae332d840afc4501c8f6a2017128fca525be4809a13", "submitted_at": "2024-10-17T17:50:38Z", "updated_at": "2025-05-24T03:55:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.13825v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.13825", "identifiers": {"arxiv": "2410.13825"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13825"}], "name": "AgentOccam: A Simple Yet Strong Baseline for LLM-Based Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies how simplifying and aligning a web agent's observations and actions with language-model capabilities affects task completion.", "topics": ["web-navigation", "planning"], "updated_at": "2025-05-24", "url": "https://arxiv.org/abs/2410.13825"}
{"aliases": [], "authors": ["Priyanshu Kumar", "Elaine Lau", "Saranya Vijayakumar", "Tu Trinh", "Scale Red Team", "Elaine Chang", "Vaughn Robinson", "Sean Hendryx", "Shuyan Zhou", "Matt Fredrikson", "Summer Yue", "Zifan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4922d9362b23bc0e41d47dacf590c6e9bd6e88cfa6418cf3517814131295270c", "submitted_at": "2024-10-11T06:54:12Z", "updated_at": "2024-10-21T18:06:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.13886v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.13886", "identifiers": {"arxiv": "2410.13886"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13886"}], "name": "Refusal-Trained LLMs Are Easily Jailbroken As Browser Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces BrowserART and studies whether chat-model refusal behavior transfers to browser agents carrying out potentially harmful instructions.", "topics": ["safety-security", "evaluation"], "updated_at": "2024-10-21", "url": "https://arxiv.org/abs/2410.13886"}
{"aliases": [], "authors": ["Taiyi Wang", "Zhihao Wu", "Jianheng Liu", "Jianye Hao", "Jun Wang", "Kun Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9e3d8350cd0be5d4f3e82b3b35a15fe89403cd602e8bb20b7cd783c7a7067a0", "submitted_at": "2024-10-18T18:19:56Z", "updated_at": "2025-02-21T16:23:59Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2410.14803v5"}, "publication": {"authors": ["Taiyi Wang", "Zhihao Wu", "Jianheng Liu", "Jianye HAO", "Jun Wang", "Kun Shao"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher title uses singular Agent; the preprint uses Agents. The authors and DistRL contribution agree.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "74757-74782", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/b9e472cd579c83e2f6aa3459f46aac28-Paper-Conference.pdf", "publisher": null, "series": null, "title": "DistRL: An Asynchronous Distributed Reinforcement Learning Framework for On-Device Control Agent", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "f3c834388e958e41ea7a8419b354435c1c522ef256c00d3af7f72bebb89c7fdc", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.14803", "identifiers": {"arxiv": "2410.14803"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.14803"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/b9e472cd579c83e2f6aa3459f46aac28-Paper-Conference.pdf"}], "name": "DistRL: An Asynchronous Distributed Reinforcement Learning Framework for On-Device Control Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces asynchronous distributed reinforcement learning with centralized updates and decentralized interaction collection for device-control agents.", "topics": ["reinforcement-learning", "efficiency"], "updated_at": "2025-02-21", "url": "https://arxiv.org/abs/2410.14803"}
{"aliases": [], "authors": ["Filippo Iulianelli", "Sung Kim", "Joshua Sussan", "Aaron D. Lauda"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "954b0b115dca3a02c90a4c08fd5b6b8a21960ba4ee138839840baba95ff63d3f", "submitted_at": "2024-10-18T21:03:07Z", "updated_at": "2025-04-26T18:57:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.14860v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.14860", "identifiers": {"arxiv": "2410.14860"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.14860"}], "name": "Universal quantum computation using Ising anyons from a non-semisimple Topological Quantum Field Theory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-26", "url": "https://arxiv.org/abs/2410.14860"}
{"aliases": [], "authors": ["Jingxuan Chen", "Derek Yuen", "Bin Xie", "Yuhao Yang", "Gongwei Chen", "Zhihao Wu", "Li Yixing", "Xurui Zhou", "Weiwen Liu", "Shuai Wang", "Kaiwen Zhou", "Rui Shao", "Liqiang Nie", "Yasheng Wang", "Jianye Hao", "Jun Wang", "Kun Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "206059754652205f1e036fac8ee117f71dd356bfe2d35fd84aff8d2935ea242d", "submitted_at": "2024-10-19T17:28:48Z", "updated_at": "2025-03-31T20:39:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.15164v3"}, "publication": {"authors": ["Jingxuan Chen", "Derek Yuen", "Bin Xie", "Yuhao Yang", "Gongwei Chen", "Zhihao Wu", "Li Yixing", "Xurui Zhou", "Weiwen Liu", "Shuai Wang", "Kaiwen Zhou", "Rui Shao", "Liqiang Nie", "Yasheng Wang", "Jianye HAO", "Jun Wang", "Kun Shao"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "61769-61806", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/9a75f4dd9679aa4ff80a4e6f0a1dc700-Paper-Conference.pdf", "publisher": null, "series": null, "title": "SPA-BENCH: A COMPREHENSIVE BENCHMARK FOR SMARTPHONE AGENT EVALUATION", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "56efdfea12b96edd94ad366f941d57668cdb75f6889fc03297e749497ff03e92", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.15164", "identifiers": {"arxiv": "2410.15164"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.15164"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/9a75f4dd9679aa4ff80a4e6f0a1dc700-Paper-Conference.pdf"}], "name": "SPA-Bench: A Comprehensive Benchmark for SmartPhone Agent Evaluation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a smartphone-agent benchmark with English and Chinese tasks, real-time Android interaction, and completion and resource-use metrics.", "topics": ["evaluation", "efficiency", "mobile-navigation"], "updated_at": "2025-03-31", "url": "https://arxiv.org/abs/2410.15164"}
{"aliases": [], "authors": ["Philip Baillargeon", "Jina Yoon", "Amy Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eca3e724763c6596cbd9a2ebf325b5bf2e2162ef0a6be38e28e4c05781ab325a", "submitted_at": "2024-10-20T22:43:11Z", "updated_at": "2024-10-20T22:43:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.15527v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.15527", "identifiers": {"arxiv": "2410.15527"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.15527"}], "name": "Who Puts the \"Social\" in \"Social Computing\"?: Using A Neurodiversity Framing to Review Social Computing Research", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-20", "url": "https://arxiv.org/abs/2410.15527"}
{"aliases": [], "authors": ["ByungOk Han", "Jaehong Kim", "Jinhyeok Jang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "41a0d429d88819301715374cd95b5a8e7d758ec1da32a2a5c7ccf8f9b29f1647", "submitted_at": "2024-10-21T00:36:02Z", "updated_at": "2024-10-21T00:36:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.15549v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.15549", "identifiers": {"arxiv": "2410.15549"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.15549"}], "name": "A Dual Process VLA: Efficient Robotic Manipulation Leveraging VLM", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-21", "url": "https://arxiv.org/abs/2410.15549"}
{"aliases": [], "authors": ["Yueqi Song", "Frank Xu", "Shuyan Zhou", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0944387dd3c65ea7b2f5ca52721496cf10b3c9860b303063201e7811318f4950", "submitted_at": "2024-10-21T19:46:06Z", "updated_at": "2025-06-16T22:23:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.16464v3"}, "publication": {"authors": ["Yueqi Song", "Frank F. Xu", "Shuyan Zhou", "Graham Neubig"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.577", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.577/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "11066-11085", "pdf_url": "https://aclanthology.org/2025.findings-acl.577.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Beyond Browsing: API-Based Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.577/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:55Z", "role": "primary", "sha256": "12c4aaa6eb73e55e905fb507db47a14cc245c933504453365b06242c39e92481", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.577/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.16464", "identifiers": {"arxiv": "2410.16464"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.16464"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.577/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.577.pdf"}], "name": "Beyond Browsing: API-Based Web Agents", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-10-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares browsing-only, API-only, and hybrid web agents on WebArena, examining how access to application APIs changes task execution alongside browser interaction.", "topics": ["web-navigation", "tool-integration", "evaluation"], "updated_at": "2025-06-16", "url": "https://arxiv.org/abs/2410.16464"}
{"aliases": [], "authors": ["Itay Nakash", "George Kour", "Guy Uziel", "Ateret Anaby-Tavor"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "46cfc7ff0910c5406b50f39ed4458bdc6fac4232358665f02801f8d66e09c76b", "submitted_at": "2024-10-22T12:24:41Z", "updated_at": "2024-10-22T12:24:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.16950v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.16950", "identifiers": {"arxiv": "2410.16950"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.16950"}], "name": "Breaking ReAct Agents: Foot-in-the-Door Attack Will Get You In", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-22", "url": "https://arxiv.org/abs/2410.16950"}
{"aliases": [], "authors": ["Jiaxuan Zhang", "Zhao-Yun Chen", "Jia-Ning Li", "Tian-Hao Wei", "Huan-Yu Liu", "Xi-Ning Zhuang", "Qing-Song Li", "Yu-Chun Wu", "Guo-Ping Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7877e0507a3be6dc36cd543d484e24304715b6ebdeb6349cd7345d2d7f2cc595", "submitted_at": "2024-10-22T12:44:41Z", "updated_at": "2025-01-02T09:04:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.16963v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.16963", "identifiers": {"arxiv": "2410.16963"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.16963"}], "name": "Scalable Constant-Time Logical Gates for Large-Scale Quantum Computation Using Window-Based Correlated Decoding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-02", "url": "https://arxiv.org/abs/2410.16963"}
{"aliases": [], "authors": ["Hongru Cai", "Yongqi Li", "Wenjie Wang", "Fengbin Zhu", "Xiaoyu Shen", "Wenjie Li", "Tat-Seng Chua"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08321196802dd2784d3bd45df49e6659de09e00c98319cee83107d4978d7ce7b", "submitted_at": "2024-10-22T17:54:45Z", "updated_at": "2025-03-24T17:51:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.17236v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.17236", "identifiers": {"arxiv": "2410.17236"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.17236"}], "name": "Large Language Models Empowered Personalized Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-24", "url": "https://arxiv.org/abs/2410.17236"}
{"aliases": [], "authors": ["Chejian Xu", "Mintong Kang", "Jiawei Zhang", "Zeyi Liao", "Lingbo Mo", "Mengqi Yuan", "Huan Sun", "Bo Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "226a4acbbaab42c8cbc44a897651d8c745119426ea1223f1b849558af32fab02", "submitted_at": "2024-10-22T20:18:26Z", "updated_at": "2025-05-31T18:34:01Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2410.17401v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.17401", "identifiers": {"arxiv": "2410.17401"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.17401"}], "name": "AdvAgent: Controllable Blackbox Red-teaming on Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a black-box web-agent red-teaming framework that learns adversarial prompts from agent feedback to study security weaknesses.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-05-31", "url": "https://arxiv.org/abs/2410.17401"}
{"aliases": [], "authors": ["Juyong Lee", "Dongyoon Hahm", "June Suk Choi", "W. Bradley Knox", "Kimin Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "362cbf7ea267ddc4a64505aa95c4f16a9f3631f95319ce10eef9981e2c85937b", "submitted_at": "2024-10-23T02:51:43Z", "updated_at": "2026-01-27T01:51:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.17520v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.17520", "identifiers": {"arxiv": "2410.17520"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.17520"}], "name": "MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-10-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates mobile agents on misuse, side effects, and indirect prompt-injection risks in an Android emulator environment.", "topics": ["safety-security", "evaluation"], "updated_at": "2026-01-27", "url": "https://arxiv.org/abs/2410.17520"}
{"aliases": [], "authors": ["Filippos Christianos", "Georgios Papoudakis", "Thomas Coste", "Jianye Hao", "Jun Wang", "Kun Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9f6eb76587ef91a21942d34ba9ee7ba0ebc846277b12c69ecb179c2b2d1720df", "submitted_at": "2024-10-23T13:57:00Z", "updated_at": "2025-02-12T17:51:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.17883v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.17883", "identifiers": {"arxiv": "2410.17883"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.17883"}], "name": "Lightweight Neural App Control", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "LiMAC combines a compact Action Transformer with a fine-tuned vision-language model for Android app control. It uses textual goals and histories of screenshots and UI trees to predict actions, with evaluation on two mobile-control datasets.", "topics": ["mobile-navigation", "action-prediction", "efficiency", "interaction-history"], "updated_at": "2025-02-12", "url": "https://arxiv.org/abs/2410.17883"}
{"aliases": [], "authors": ["Chengyou Jia", "Minnan Luo", "Zhuohang Dang", "Qiushi Sun", "Fangzhi Xu", "Junlin Hu", "Tianbao Xie", "Zhiyong Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ece9f64cc037dad8633388edf1b54d88818d8092b77e19b15780e929b23aa593", "submitted_at": "2024-10-24T09:58:40Z", "updated_at": "2024-10-24T09:58:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.18603v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.18603", "identifiers": {"arxiv": "2410.18603"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.18603"}], "name": "AgentStore: Scalable Integration of Heterogeneous Agents As Specialized Generalist Computer Assistant", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a platform that integrates specialized agents under a coordinating MetaAgent for varied computer tasks.", "topics": ["automation", "tool-integration"], "updated_at": "2024-10-24", "url": "https://arxiv.org/abs/2410.18603"}
{"aliases": [], "authors": ["Zonghao Ying", "Aishan Liu", "Siyuan Liang", "Lei Huang", "Jinyang Guo", "Wenbo Zhou", "Xianglong Liu", "Dacheng Tao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ef7c05ccfb2a473c9a209f169240c96301912192fda971317abe9059057aaebc", "submitted_at": "2024-10-24T17:14:40Z", "updated_at": "2024-10-24T17:14:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.18927v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2410.18927", "identifiers": {"arxiv": "2410.18927"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.18927"}], "name": "SafeBench: A Safety Evaluation Framework for Multimodal Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-24", "url": "https://arxiv.org/abs/2410.18927"}
{"aliases": [], "authors": ["Xiaoqiang Wang", "Bang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "66fa51b108ac79cc43a895ba757ca463d5008bfe6db5cde4bdf3a3bb38c444eb", "submitted_at": "2024-10-24T17:58:08Z", "updated_at": "2024-10-24T17:58:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.18963v1"}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.18963", "identifiers": {"arxiv": "2410.18963"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.18963"}], "name": "OSCAR: Operating System Control via State-Aware Reasoning and Re-Planning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "OSCAR translates user instructions into Python control code while observing screen images and acting through mouse and keyboard operations. A state-machine controller handles errors and replans tasks during execution across desktop and mobile applications.", "topics": ["planning", "state-tracking", "tool-integration", "robustness"], "updated_at": "2024-10-24", "url": "https://arxiv.org/abs/2410.18963"}
{"aliases": [], "authors": ["Zhangheng Li", "Keen You", "Haotian Zhang", "Di Feng", "Harsh Agrawal", "Xiujun Li", "Mohana Prasad Sathya Moorthy", "Jeff Nichols", "Yinfei Yang", "Zhe Gan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "299b7c631468e259ac9d25a1c4ebf507be57b7ee478c611d0a2597b588f6c4c8", "submitted_at": "2024-10-24T17:58:31Z", "updated_at": "2025-02-28T00:29:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.18967v2"}}, "environments": ["ios", "android", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.18967", "identifiers": {"arxiv": "2410.18967"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.18967"}], "name": "Ferret-UI 2: Mastering Universal User Interface Understanding Across Platforms", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Extends Ferret-UI across device platforms with adaptive visual scaling and synthetic data for referring, grounding, and user-centered UI tasks.", "topics": ["screen-understanding", "grounding"], "updated_at": "2025-02-28", "url": "https://arxiv.org/abs/2410.18967"}
{"aliases": [], "authors": ["Revanth Gangi Reddy", "Sagnik Mukherjee", "Jeonghwan Kim", "Zhenhailong Wang", "Dilek Hakkani-Tur", "Heng Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b9938c5caccd6d0d76e817b0ba1e9f174a203b8b9f6d6d562b0079c2f2ed103", "submitted_at": "2024-10-24T18:01:28Z", "updated_at": "2024-10-24T18:01:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.19054v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.19054", "identifiers": {"arxiv": "2410.19054"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.19054"}], "name": "Infogent: An Agent-Based Framework for Web Information Aggregation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-24", "url": "https://arxiv.org/abs/2410.19054"}
{"aliases": [], "authors": ["Lawrence Jang", "Yinheng Li", "Dan Zhao", "Charles Ding", "Justin Lin", "Paul Pu Liang", "Rogerio Bonatti", "Kazuhito Koishida"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f413943bd5c2a395ddb5cf38a1a5ad61754b35a9eb5de43e8d980bd83e4adff", "submitted_at": "2024-10-24T19:03:01Z", "updated_at": "2025-02-15T05:19:38Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.19100v3"}, "publication": {"authors": ["Lawrence Jang", "Yinheng Li", "Dan Zhao", "Charles Ding", "Justin Lin", "Paul Pu Liang", "Rogerio Bonatti", "Kazuhito Koishida"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "36934-36958", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/5b555804d495321df2e3208cc27f4fbc-Paper-Conference.pdf", "publisher": null, "series": null, "title": "VideoWebArena:  Evaluating Long Context Multimodal Agents with Video Understanding Web Tasks", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "fdb42289bd00ad80cb51b66a01b91e13e60f134e809bb6c46181b4024ccabad2", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.19100", "identifiers": {"arxiv": "2410.19100"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.19100"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/5b555804d495321df2e3208cc27f4fbc-Paper-Conference.pdf"}], "name": "VideoWebArena: Evaluating Long Context Multimodal Agents with Video Understanding Web Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates whether web agents can retain procedural skills and factual information from long video tutorials and use them during tasks.", "topics": ["evaluation", "video", "memory"], "updated_at": "2025-02-15", "url": "https://arxiv.org/abs/2410.19100"}
{"aliases": [], "authors": ["Xuetian Chen", "Hangcheng Li", "Jiaqing Liang", "Sihang Jiang", "Deqing Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dce2e22787c6e2cce95ecc4f5583bf94d7670b4b87ab75d2b8c7efb66fbcd1a8", "submitted_at": "2024-10-25T10:46:17Z", "updated_at": "2024-11-02T08:54:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.19461v2"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.19461", "identifiers": {"arxiv": "2410.19461"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.19461"}], "name": "EDGE: Enhanced Grounded GUI Understanding with Enriched Multi-Granularity Synthetic Data", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Synthesizes GUI training examples at multiple levels of detail from webpages and studies transfer of learned understanding to other interface environments.", "topics": ["data-generation", "grounding", "screen-understanding"], "updated_at": "2024-11-02", "url": "https://arxiv.org/abs/2410.19461"}
{"aliases": [], "authors": ["Hongliang He", "Wenlin Yao", "Kaixin Ma", "Wenhao Yu", "Hongming Zhang", "Tianqing Fang", "Zhenzhong Lan", "Dong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db2a115281f831ea9333a03fc2a9ebb2f09528eb082060f00a33b125f9ef7762", "submitted_at": "2024-10-25T15:01:27Z", "updated_at": "2024-10-25T15:01:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.19609v1"}, "publication": {"authors": ["Hongliang He", "Wenlin Yao", "Kaixin Ma", "Wenhao Yu", "Hongming Zhang", "Tianqing Fang", "Zhenzhong Lan", "Dong Yu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.1336", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.1336/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "27545-27564", "pdf_url": "https://aclanthology.org/2025.acl-long.1336.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "OpenWebVoyager: Building Multimodal Web Agents via Iterative Real-World Exploration, Feedback and Optimization", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.1336/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:55Z", "role": "primary", "sha256": "efd9fe9277a980d42a991abe428ccc3983299c6a8a4d91e18d19cc07a37ed22c", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.1336/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.19609", "identifiers": {"arxiv": "2410.19609"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.19609"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.1336/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.1336.pdf"}], "name": "OpenWebVoyager: Building Multimodal Web Agents via Iterative Real-World Exploration, Feedback and Optimization", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Develops multimodal browser agents through repeated cycles of real-web exploration, trajectory feedback, and policy optimization.", "topics": ["web-navigation", "data-generation"], "updated_at": "2024-10-25", "url": "https://arxiv.org/abs/2410.19609"}
{"aliases": [], "authors": ["Yilun Jin", "Zheng Li", "Chenwei Zhang", "Tianyu Cao", "Yifan Gao", "Pratik Jayarao", "Mao Li", "Xin Liu", "Ritesh Sarkhel", "Xianfeng Tang", "Haodong Wang", "Zhengyang Wang", "Wenju Xu", "Jingfeng Yang", "Qingyu Yin", "Xian Li", "Priyanka Nigam", "Yi Xu", "Kai Chen", "Qiang Yang", "Meng Jiang", "Bing Yin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3b37256be47d7c8260d232f358dc92c316670f938cc9b0658c9f47a4aec0b3e1", "submitted_at": "2024-10-28T05:25:47Z", "updated_at": "2024-10-31T12:54:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.20745v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.20745", "identifiers": {"arxiv": "2410.20745"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.20745"}], "name": "Shopping MMLU: A Massive Multi-Task Online Shopping Benchmark for Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-31", "url": "https://arxiv.org/abs/2410.20745"}
{"aliases": [], "authors": ["Xiao Xia", "Dan Zhang", "Zibo Liao", "Zhenyu Hou", "Tianrui Sun", "Jing Li", "Ling Fu", "Yuxiao Dong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0945222669ea174806a5f3d951d81ef4925eb264068e01ded866024fb479df11", "submitted_at": "2024-10-29T10:01:40Z", "updated_at": "2025-06-26T06:24:08Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.21909v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.21909", "identifiers": {"arxiv": "2410.21909"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.21909"}], "name": "SceneGenAgent: Precise Industrial Scene Generation with Coding Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-26", "url": "https://arxiv.org/abs/2410.21909"}
{"aliases": [], "authors": ["Jaekyeom Kim", "Dong-Ki Kim", "Lajanugen Logeswaran", "Sungryull Sohn", "Honglak Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf1fb8aa2e289a72f3c44c99e2d531eaa5eeea08b78ff6f3a21e139d47502b17", "submitted_at": "2024-10-29T21:37:04Z", "updated_at": "2024-10-29T21:37:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.22552v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.22552", "identifiers": {"arxiv": "2410.22552"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.22552"}], "name": "Auto-Intent: Automated Intent Discovery and Self-Exploration for Large Language Model Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Discovers compact intents from demonstrations and uses intent predictions to guide a language-model web agent's decisions and exploration.", "topics": ["planning", "web-navigation"], "updated_at": "2024-10-29", "url": "https://arxiv.org/abs/2410.22552"}
{"aliases": [], "authors": ["Natasha Butt", "Varun Chandrasekaran", "Neel Joshi", "Besmira Nushi", "Vidhisha Balachandran"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6577b6af1aa01a0a78e387c2728e2323f867e746617119c8ef0f1050835cc529", "submitted_at": "2024-10-29T22:56:18Z", "updated_at": "2025-10-07T07:17:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.22584v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.22584", "identifiers": {"arxiv": "2410.22584"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.22584"}], "name": "BenchAgents: Multi-Agent Systems for Structured Benchmark Creation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-07", "url": "https://arxiv.org/abs/2410.22584"}
{"aliases": [], "authors": ["Hao Li", "Xiaogeng Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a939e7fb30d9d4e90c548c1f72c8930fb54daab579c848213a0afd61f9aacba", "submitted_at": "2024-10-30T07:39:42Z", "updated_at": "2025-03-30T16:39:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.22770v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.22770", "identifiers": {"arxiv": "2410.22770"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.22770"}], "name": "InjecGuard: Benchmarking and Mitigating Over-defense in Prompt Injection Guardrail Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-30", "url": "https://arxiv.org/abs/2410.22770"}
{"aliases": [], "authors": ["Qinqing Zheng", "Mikael Henaff", "Amy Zhang", "Aditya Grover", "Brandon Amos"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2151e587e16c83a883e23aa57aaee2ce62a1a719ae8ed7ed4dbf0267051be9b8", "submitted_at": "2024-10-30T13:52:43Z", "updated_at": "2025-10-23T23:54:03Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2410.23022v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2410.23022", "identifiers": {"arxiv": "2410.23022"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.23022"}], "name": "Online Intrinsic Rewards for Decision Making Agents from Large Language Model Feedback", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-23", "url": "https://arxiv.org/abs/2410.23022"}
{"aliases": [], "authors": ["Zhiyong Wu", "Zhenyu Wu", "Fangzhi Xu", "Yian Wang", "Qiushi Sun", "Chengyou Jia", "Kanzhi Cheng", "Zichen Ding", "Liheng Chen", "Paul Pu Liang", "Yu Qiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e0dbd8ef8807e7886c02ed6296c9d188503fe7b676062fcff1890fd0652f10f3", "submitted_at": "2024-10-30T17:10:19Z", "updated_at": "2024-10-30T17:10:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.23218v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.23218", "identifiers": {"arxiv": "2410.23218"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.23218"}], "name": "OS-ATLAS: A Foundation Action Model for Generalist GUI Agents", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-10-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines cross-platform GUI grounding data synthesis with action-model training to study generalization to unfamiliar interfaces.", "topics": ["grounding", "data-generation"], "updated_at": "2024-10-30", "url": "https://arxiv.org/abs/2410.23218"}
{"aliases": [], "authors": ["Haoyi Qiu", "Alexander R. Fabbri", "Divyansh Agarwal", "Kung-Hsiang Huang", "Sarah Tan", "Nanyun Peng", "Chien-Sheng Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "26fac89c78ab03fcbaf25bbfe93ebce090dec9adbf5c6dd5440641f154cffac9", "submitted_at": "2024-10-30T17:35:44Z", "updated_at": "2025-03-08T23:37:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.23252v3"}, "publication": {"authors": ["Haoyi Qiu", "Alexander Richard Fabbri", "Divyansh Agarwal", "Kung-Hsiang Huang", "Sarah Tan", "Nanyun Peng", "Chien-Sheng Wu"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.findings-naacl.222", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-naacl.222/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "3978-4005", "pdf_url": "https://aclanthology.org/2025.findings-naacl.222.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Evaluating Cultural and Social Awareness of LLM Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-naacl.222/", "venue": "Findings of the Association for Computational Linguistics: NAACL 2025", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:55Z", "role": "primary", "sha256": "6563b807356c30b69a1c5bb9328358c139d171777cb893079ddc123cb82a282e", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-naacl.222/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.23252", "identifiers": {"arxiv": "2410.23252"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.23252"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-naacl.222/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-naacl.222.pdf"}], "name": "Evaluating Cultural and Social Awareness of LLM Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CASA to assess web agents' responses to cultural and social norms in shopping and forum tasks, including misleading observations.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-03-08", "url": "https://arxiv.org/abs/2410.23252"}
{"aliases": [], "authors": ["Nalin Tiwary", "Vardhan Dongre", "Sanil Arun Chawla", "Ashwin Lamani", "Dilek Hakkani-Tür"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c67b1559410152607fc9d54abf25c37ef939c9f755f089bdf87d4f39393c085", "submitted_at": "2024-10-31T01:51:41Z", "updated_at": "2024-10-31T01:51:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.23555v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:27Z", "role": "primary", "sha256": "fced2234354232b8834eef0cf2557b2932990e74b3530fc95459cd3fa0e11792", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2410.23555v1"}], "id": "arxiv:2410.23555", "identifiers": {"arxiv": "2410.23555"}, "kind": "paper", "last_verified_at": "2026-10-17T09:13:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.23555"}], "name": "From Context to Action: Analysis of the Impact of State Representation and Context on the Generalization of Multi-Turn Web Navigation Agents", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-10-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes how interaction history and webpage representation affect multiturn web-agent generalization across unfamiliar websites, task categories, and geographic settings.", "topics": ["web-navigation", "interaction-history", "state-tracking"], "updated_at": "2024-10-31", "url": "https://arxiv.org/abs/2410.23555"}
{"aliases": [], "authors": ["Cihan Okay", "Atak Talay Yucel", "Selman Ipek"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c2e51fcb2f87b288b79ed479a74e157ffed0e45d35e21197bd3d9c2b0f253e5", "submitted_at": "2024-10-31T08:40:51Z", "updated_at": "2024-10-31T08:40:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.23734v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2410.23734", "identifiers": {"arxiv": "2410.23734"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2410.23734"}], "name": "Classical simulation of universal measurement-based quantum computation using multipartite Bell scenarios", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-10-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-10-31", "url": "https://arxiv.org/abs/2410.23734"}
{"aliases": [], "authors": ["Yifan Xu", "Xiao Liu", "Xueqiao Sun", "Siyi Cheng", "Hao Yu", "Hanyu Lai", "Shudan Zhang", "Dan Zhang", "Jie Tang", "Yuxiao Dong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d29fe5262adf8677bbd5fead628c5fc28707af263aa7e800d1dee1f7e499784c", "submitted_at": "2024-10-31T15:25:20Z", "updated_at": "2024-11-04T05:57:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.24024v2"}, "publication": {"authors": ["Yifan Xu", "Xiao Liu", "Xueqiao Sun", "Siyi Cheng", "Hao Yu (余浩)", "Hanyu Lai", "Shudan Zhang", "Dan Zhang", "Jie Tang", "Yuxiao Dong"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.107", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.107/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "2144-2166", "pdf_url": "https://aclanthology.org/2025.acl-long.107.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.107/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:56Z", "role": "primary", "sha256": "52b82854b5f4703e6e7f3cd3e9fc49adcd011ca3ba0a8fd1aced77214c452043", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.107/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2410.24024", "identifiers": {"arxiv": "2410.24024"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.24024"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.107/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.107.pdf"}], "name": "AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agents", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-10-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Provides a shared Android action environment, reproducible tasks, and instruction data for training and comparing language and multimodal agents.", "topics": ["evaluation", "data-generation", "mobile-navigation"], "updated_at": "2024-11-04", "url": "https://arxiv.org/abs/2410.24024"}
{"aliases": [], "authors": ["Xiao Liu", "Bo Qin", "Dongzhu Liang", "Guang Dong", "Hanyu Lai", "Hanchen Zhang", "Hanlin Zhao", "Iat Long Iong", "Jiadai Sun", "Jiaqi Wang", "Junjie Gao", "Junjun Shan", "Kangning Liu", "Shudan Zhang", "Shuntian Yao", "Siyi Cheng", "Wentao Yao", "Wenyi Zhao", "Xinghan Liu", "Xinyi Liu", "Xinying Chen", "Xinyue Yang", "Yang Yang", "Yifan Xu", "Yu Yang", "Yujia Wang", "Yulin Xu", "Zehan Qi", "Yuxiao Dong", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cfe1aa31818f06567aa3d619ad90ac4bdb9f9f9003a1a8a71294cd4f8ccf8dc", "submitted_at": "2024-10-28T17:05:10Z", "updated_at": "2024-10-28T17:05:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.00820v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2411.00820", "identifiers": {"arxiv": "2411.00820"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.00820"}], "name": "AutoGLM: Autonomous Foundation Agents for GUIs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-10-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents AutoGLM agents for browser and phone control, separating planning from grounding and training through an evolving interaction curriculum.", "topics": ["automation", "reinforcement-learning", "grounding"], "updated_at": "2024-10-28", "url": "https://arxiv.org/abs/2411.00820"}
{"aliases": [], "authors": ["Biao Wu", "Yanda Li", "Zhiwei Zhang", "Yunchao Wei", "Meng Fang", "Ling Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de860a29f6c54317586c2218688436898f1d7a8b61207b1d5f4a65b05b3a4e6e", "submitted_at": "2024-11-04T11:50:58Z", "updated_at": "2025-09-14T09:13:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2411.02006v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:27Z", "role": "primary", "sha256": "bda47fcb119c581a798dbc791b4929af91832d5053ac073a02dab8fd4e10ace3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2411.02006v3"}], "id": "arxiv:2411.02006", "identifiers": {"arxiv": "2411.02006"}, "kind": "survey", "last_verified_at": "2026-10-17T09:13:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.02006"}], "name": "Foundations and Recent Trends in Multimodal Mobile Agents: A Survey", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-11-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Surveys multimodal mobile agents, organizing prompt-based and training-based approaches alongside evaluation benchmarks, supporting technologies, and open challenges in mobile task execution.", "topics": ["survey", "mobile-navigation", "evaluation"], "updated_at": "2025-09-14", "url": "https://arxiv.org/abs/2411.02006"}
{"aliases": [], "authors": ["Zehan Qi", "Xiao Liu", "Iat Long Iong", "Hanyu Lai", "Xueqiao Sun", "Wenyi Zhao", "Yu Yang", "Xinyue Yang", "Jiadai Sun", "Shuntian Yao", "Tianjie Zhang", "Wei Xu", "Jie Tang", "Yuxiao Dong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db039a2592cd79a1f8f7f1c985c2ddf42447c63d1f58d2815ba1d6ef2369da47", "submitted_at": "2024-11-04T17:59:58Z", "updated_at": "2025-01-27T11:56:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2411.02337v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.02337", "identifiers": {"arxiv": "2411.02337"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.02337"}], "name": "WebRL: Training LLM Web Agents via Self-Evolving Online Curriculum Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-11-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains web agents with an evolving task curriculum, outcome supervision, and reinforcement-learning updates intended to manage policy drift.", "topics": ["reinforcement-learning", "web-navigation"], "updated_at": "2025-01-27", "url": "https://arxiv.org/abs/2411.02337"}
{"aliases": [], "authors": ["Yanzhe Zhang", "Tao Yu", "Diyi Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "085c64595365cc6439c8c5fa8a479259a3d59b7e9740e0f9d43e48d00e1439c0", "submitted_at": "2024-11-04T18:56:42Z", "updated_at": "2025-05-24T16:15:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.02391v2"}, "publication": {"authors": ["Yanzhe Zhang", "Tao Yu", "Diyi Yang"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.411", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.411/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "8387-8401", "pdf_url": "https://aclanthology.org/2025.acl-long.411.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Attacking Vision-Language Computer Agents via Pop-ups", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.411/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:56Z", "role": "primary", "sha256": "a28a603d74e08954a05265f54c8765bd874a6cb2fc0c5a68a35e15e68692b0a1", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.411/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2411.02391", "identifiers": {"arxiv": "2411.02391"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.02391"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.411/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.411.pdf"}], "name": "Attacking Vision-Language Computer Agents via Pop-ups", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-11-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Tests whether adversarial visual pop-ups divert computer agents from intended tasks in desktop and web benchmark environments.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-05-24", "url": "https://arxiv.org/abs/2411.02391"}
{"aliases": [], "authors": ["Shivam Duggal", "Phillip Isola", "Antonio Torralba", "William T. Freeman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d35423688cdf40c74477250a0e296d7ea47efb0829a9444cab252cb50cfec498", "submitted_at": "2024-11-04T18:58:01Z", "updated_at": "2024-11-04T18:58:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.02393v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2411.02393", "identifiers": {"arxiv": "2411.02393"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.02393"}], "name": "Adaptive Length Image Tokenization via Recurrent Allocation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-04", "url": "https://arxiv.org/abs/2411.02393"}
{"aliases": [], "authors": ["Mike Lambeta", "Tingfan Wu", "Ali Sengul", "Victoria Rose Most", "Nolan Black", "Kevin Sawyer", "Romeo Mercado", "Haozhi Qi", "Alexander Sohn", "Byron Taylor", "Norb Tydingco", "Gregg Kammerer", "Dave Stroud", "Jake Khatha", "Kurt Jenkins", "Kyle Most", "Neal Stein", "Ricardo Chavira", "Thomas Craven-Bartle", "Eric Sanchez", "Yitian Ding", "Jitendra Malik", "Roberto Calandra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "25589000fb32af2b1ac0d8dea77400f594d2fffda2d7eb26dd4e5abee7413a03", "submitted_at": "2024-11-04T18:38:50Z", "updated_at": "2024-11-04T18:38:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.02479v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2411.02479", "identifiers": {"arxiv": "2411.02479"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.02479"}], "name": "Digitizing Touch with an Artificial Multimodal Fingertip", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-04", "url": "https://arxiv.org/abs/2411.02479"}
{"aliases": [], "authors": ["Kevin Y. Li", "Sachin Goyal", "Joao D. Semedo", "J. Zico Kolter"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e0a56fd620c3db0d18730447159d0a1a3d1de214ff35141f3b61f10914820a38", "submitted_at": "2024-11-05T18:54:21Z", "updated_at": "2025-04-21T09:34:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.03312v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2411.03312", "identifiers": {"arxiv": "2411.03312"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.03312"}], "name": "Inference Optimal VLMs Need Fewer Visual Tokens and More Parameters", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-21", "url": "https://arxiv.org/abs/2411.03312"}
{"aliases": [], "authors": ["Zhirui Deng", "Zhicheng Dou", "Yutao Zhu", "Ji-Rong Wen", "Ruibin Xiong", "Mang Wang", "Weipeng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd631a31f72dabe3a3bebf322031f9d686082073a8e4aa71eb65745498327f4e", "submitted_at": "2024-11-06T10:35:11Z", "updated_at": "2024-12-09T09:20:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2411.03817v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:28Z", "role": "primary", "sha256": "b105dc6cc1b32ab71467b2701b6a85506f698e73f04cfbd63ab41eddfbca48b6", "supports": ["name", "kind", "summary", "topics", "notes"], "url": "https://arxiv.org/abs/2411.03817v3"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:31Z", "role": "primary", "sha256": "ce801ec21577b4f62d5b0ecec7c50e87744e5b38db612a2aa4017e27fe37102f", "supports": ["summary", "environments", "scope", "notes"], "url": "https://arxiv.org/html/2411.03817v3"}], "id": "arxiv:2411.03817", "identifiers": {"arxiv": "2411.03817"}, "kind": "paper", "last_verified_at": "2026-10-17T09:13:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.03817"}], "name": "From Novice to Expert: LLM Agent Policy Optimization via Step-wise Reinforcement Learning", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced. Official full-paper experimental sections explicitly evaluate WebShop and Mind2Web; the overall optimization method also covers embodied and question-answering tasks.", "published_at": "2024-11-06", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces StepAgent, which compares agent and expert actions to supply intermediate rewards for policy learning, with implicit-reward and inverse-RL variants. Evaluation includes WebShop and Mind2Web alongside non-web tasks.", "topics": ["reinforcement-learning", "reward-models", "self-reflection"], "updated_at": "2024-12-09", "url": "https://arxiv.org/abs/2411.03817"}
{"aliases": [], "authors": ["Adam Fourney", "Gagan Bansal", "Hussein Mozannar", "Cheng Tan", "Eduardo Salinas", "Erkang", "Zhu", "Friederike Niedtner", "Grace Proebsting", "Griffin Bassman", "Jack Gerrits", "Jacob Alber", "Peter Chang", "Ricky Loynd", "Robert West", "Victor Dibia", "Ahmed Awadallah", "Ece Kamar", "Rafah Hosn", "Saleema Amershi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4a2c499237dbcbe1482824a2710f5185dee05c9ed66a0882d1e422f30afe056a", "submitted_at": "2024-11-07T06:36:19Z", "updated_at": "2024-11-07T06:36:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.04468v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2411.04468", "identifiers": {"arxiv": "2411.04468"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.04468"}], "name": "Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-11-07", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces an orchestrated team of browsing, file, and coding agents with progress tracking and error recovery for general digital tasks.", "topics": ["planning", "tool-integration"], "updated_at": "2024-11-07", "url": "https://arxiv.org/abs/2411.04468"}
{"aliases": [], "authors": ["Shuai Wang", "Weiwen Liu", "Jingxuan Chen", "Yuqi Zhou", "Weinan Gan", "Xingshan Zeng", "Yuhan Che", "Shuai Yu", "Xinlong Hao", "Kun Shao", "Bin Wang", "Chuhan Wu", "Yasheng Wang", "Ruiming Tang", "Jianye Hao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57670445f9b0041bfb505a195d01a9ded6234e01e44c9aa43a6aee5ebd57dd93", "submitted_at": "2024-11-07T17:28:10Z", "updated_at": "2025-02-13T10:09:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.04890v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.04890", "identifiers": {"arxiv": "2411.04890"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.04890"}], "name": "GUI Agents with Foundation Models: A Comprehensive Survey", "notes": "", "published_at": "2024-11-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Survey of (M)LLM-based GUI agents organized around their datasets and benchmarks, a unified agent framework with a taxonomy, and commercial applications.", "topics": ["survey"], "updated_at": "2025-02-13", "url": "https://arxiv.org/abs/2411.04890"}
{"aliases": [], "authors": ["Shengda Fan", "Xin Cong", "Yuepeng Fu", "Zhong Zhang", "Shuyan Zhang", "Yuanwei Liu", "Yesai Wu", "Yankai Lin", "Zhiyuan Liu", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2407fd35164680db7422e4dc57c51855ba8d244f997e5a4cfe75d8dfba8ab691", "submitted_at": "2024-11-08T09:58:02Z", "updated_at": "2024-11-08T09:58:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.05451v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2411.05451", "identifiers": {"arxiv": "2411.05451"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.05451"}], "name": "WorkflowLLM: Enhancing Workflow Orchestration Capability of Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-08", "url": "https://arxiv.org/abs/2411.05451"}
{"aliases": [], "authors": ["Yu Gu", "Kai Zhang", "Yuting Ning", "Boyuan Zheng", "Boyu Gou", "Tianci Xue", "Cheng Chang", "Sanjari Srivastava", "Yanan Xie", "Peng Qi", "Huan Sun", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ecaca6448aae5c8849d008d820d700bfa438ed0248f85be585b0b831408b9b6", "submitted_at": "2024-11-10T18:50:51Z", "updated_at": "2025-04-01T05:04:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.06559v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.06559", "identifiers": {"arxiv": "2411.06559"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.06559"}], "name": "Is Your LLM Secretly a World Model of the Internet? Model-Based Planning for Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-11-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebDreamer, which simulates candidate web actions with language-model world models and value estimates before acting.", "topics": ["world-models", "planning"], "updated_at": "2025-04-01", "url": "https://arxiv.org/abs/2411.06559"}
{"aliases": [], "authors": ["Qingyu Yin", "Chak Tou Leong", "Minjun Zhu", "Hanqi Yan", "Qiang Zhang", "Yulan He", "Wenjie Li", "Jun Wang", "Yue Zhang", "Linyi Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33883febbf6cdb7590120541b49433844dae8da3ce1e70bd2e3ab55030168e27", "submitted_at": "2024-11-12T07:54:13Z", "updated_at": "2025-07-10T15:03:47Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2411.07618v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2411.07618", "identifiers": {"arxiv": "2411.07618"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.07618"}], "name": "Constrain Alignment with Sparse Autoencoders", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-10", "url": "https://arxiv.org/abs/2411.07618"}
{"aliases": [], "authors": ["Yuyou Gan", "Yong Yang", "Zhe Ma", "Ping He", "Rui Zeng", "Yiming Wang", "Qingming Li", "Chunyi Zhou", "Songze Li", "Ting Wang", "Yunjun Gao", "Yingcai Wu", "Shouling Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a9274be626fa1d68036b3784f3093da99f2d8f1c45a2aebaff9d38e64921808", "submitted_at": "2024-11-14T15:40:04Z", "updated_at": "2024-11-14T15:40:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.09523v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2411.09523", "identifiers": {"arxiv": "2411.09523"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.09523"}], "name": "Navigating the Risks: A Survey of Security, Privacy, and Ethics Threats in LLM-Based Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-14", "url": "https://arxiv.org/abs/2411.09523"}
{"aliases": [], "authors": ["Siyuan Hu", "Mingyu Ouyang", "Difei Gao", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "255e101e263c4e3a0cb491d4c34b3e75a8f3660cec2141e3a507a30cf5514dbd", "submitted_at": "2024-11-15T16:23:52Z", "updated_at": "2024-11-15T16:23:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.10323v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.10323", "identifiers": {"arxiv": "2411.10323"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.10323"}], "name": "The Dawn of GUI Agent: A Preliminary Case Study with Claude 3.5 Computer Use", "notes": "Reviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-11-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents a preliminary study of Claude 3.5 Computer Use across desktop tasks and software, analyzing planning, action, and critique behavior and providing a framework for deploying GUI-automation models.", "topics": ["evaluation", "automation", "planning"], "updated_at": "2024-11-15", "url": "https://arxiv.org/abs/2411.10323"}
{"aliases": [], "authors": ["Minghe Gao", "Wendong Bu", "Bingchen Miao", "Yang Wu", "Yunfei Li", "Juncheng Li", "Siliang Tang", "Qi Wu", "Yueting Zhuang", "Meng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "daa9b7092616e94a1d0124d36d8b75d0401b9647ec762c963524616a92def9a5", "submitted_at": "2024-11-17T02:44:56Z", "updated_at": "2024-11-17T02:44:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.10943v1"}}, "environments": ["web", "desktop", "android", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:29:13.577033+00:00", "role": "primary", "sha256": "92f18020e481d6f42a6d35e96dff820594e9370597f5392d038e9ee908c49f1e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2411.10943v1"}], "id": "arxiv:2411.10943", "identifiers": {"arxiv": "2411.10943"}, "kind": "survey", "last_verified_at": "2026-10-18T04:29:13.577033+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.10943"}], "name": "Generalist Virtual Agents: A Survey on Autonomous Agents Across Digital Platforms", "notes": "Reviewed the complete retained arXiv abstract and exact-version official HTML sections on browser/application/OS environments, screen observations, input actions and evaluation. The HTML abstract fingerprint and ordered authors match the retained Atom entry. Literature claims and experiments were not independently reproduced.", "published_at": "2024-11-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Surveys virtual agents for browsers, applications and operating systems, organizing prior work by task type, observation and action spaces, model strategies and evaluation methods. It examines screen-based interaction and keyboard, mouse and touch control, alongside transferability and long-sequence decision-making challenges.", "topics": ["survey", "web-navigation", "mobile-navigation", "evaluation", "planning"], "updated_at": "2024-11-17", "url": "https://arxiv.org/abs/2411.10943"}
{"aliases": [], "authors": ["Chen Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "884d758794e35716904f12e4839ba8660b28cb823bde2d15646ed4d1b6ae54c8", "submitted_at": "2024-11-17T18:33:02Z", "updated_at": "2024-11-17T18:33:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.11145v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2411.11145", "identifiers": {"arxiv": "2411.11145"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.11145"}], "name": "From 2D Document Interactions into Immersive Information Experience: An Example-Based Design by Augmenting Content, Spatializing Placement, Enriching Long-Term Interactions, and Simplifying Content Creations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-17", "url": "https://arxiv.org/abs/2411.11145"}
{"aliases": [], "authors": ["Yue Zhou", "Mengcheng Lan", "Xiang Li", "Litong Feng", "Yiping Ke", "Xue Jiang", "Qingyun Li", "Xue Yang", "Wayne Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3683f7e4a1b99ebe90ffa54830976697669f7beffa596bf8d2694090e822bf3e", "submitted_at": "2024-11-16T05:12:11Z", "updated_at": "2025-05-10T15:43:29Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2411.11904v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2411.11904", "identifiers": {"arxiv": "2411.11904"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.11904"}], "name": "GeoGround: A Unified Large Vision-Language Model for Remote Sensing Visual Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-10", "url": "https://arxiv.org/abs/2411.11904"}
{"aliases": [], "authors": ["Jingxuan Wei", "Cheng Tan", "Qi Chen", "Gaowei Wu", "Siyuan Li", "Zhangyang Gao", "Linzhuang Sun", "Bihui Yu", "Ruifeng Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a82d7d9749da59c4d3be15065ba8e4ea06ec1540562295fe103c241d7bc7ae46", "submitted_at": "2024-11-18T02:58:37Z", "updated_at": "2024-11-18T02:58:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.11916v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2411.11916", "identifiers": {"arxiv": "2411.11916"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.11916"}], "name": "From Words to Structured Visuals: A Benchmark and Framework for Text-to-Diagram Generation and Editing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-18", "url": "https://arxiv.org/abs/2411.11916"}
{"aliases": [], "authors": ["Xiaohan Dan", "Eitan Geva", "Victor S. Batista"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "489518d754a9f1fee56bf8f6ebd1efe927570c5e19e9fada1e778cdfd08546c0", "submitted_at": "2024-11-18T20:41:10Z", "updated_at": "2025-02-10T23:13:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.12049v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.12049", "identifiers": {"arxiv": "2411.12049"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.12049"}], "name": "Simulating Non-Markovian Quantum Dynamics on NISQ Computers Using the Hierarchical Equations of Motion", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-10", "url": "https://arxiv.org/abs/2411.12049"}
{"aliases": [], "authors": ["Anka Reuel", "Amelia Hardy", "Chandler Smith", "Max Lamparth", "Malcolm Hardy", "Mykel J. Kochenderfer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b99d2cb1500493699af2bb7688c9b6e153e26554b2537c36e8951675c7b5671", "submitted_at": "2024-11-20T02:38:24Z", "updated_at": "2024-11-20T02:38:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.12990v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2411.12990", "identifiers": {"arxiv": "2411.12990"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.12990"}], "name": "BetterBench: Assessing AI Benchmarks, Uncovering Issues, and Establishing Best Practices", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-20", "url": "https://arxiv.org/abs/2411.12990"}
{"aliases": [], "authors": ["Yingkai Ouyang", "Yumang Jing", "Gavin K. Brennen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9bd104cedc3a3cd1d5f786fed5e7f6fefe2f21530cef3355e9d1c0c337e64b1f", "submitted_at": "2024-11-20T09:16:07Z", "updated_at": "2025-09-16T01:31:02Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2411.13142v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.13142", "identifiers": {"arxiv": "2411.13142"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.13142"}], "name": "Measurement-free code-switching for low overhead quantum computation using permutation invariant codes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-16", "url": "https://arxiv.org/abs/2411.13142"}
{"aliases": [], "authors": ["Gaurav Verma", "Rachneet Kaur", "Nishan Srishankar", "Zhen Zeng", "Tucker Balch", "Manuela Veloso"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "94d908dd89ff27434805fe7915c5a93d5c2b1d873449babc1c566c5582cca734", "submitted_at": "2024-11-20T16:54:15Z", "updated_at": "2024-11-20T16:54:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.13451v1"}, "publication": {"authors": ["Gaurav Verma", "Rachneet Kaur", "Nishan Srishankar", "Zhen Zeng", "Tucker Balch", "Manuela Veloso"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.1008", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.1008/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "20635-20651", "pdf_url": "https://aclanthology.org/2025.acl-long.1008.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AdaptAgent: Adapting Multimodal Web Agents with Few-Shot Learning from Human Demonstrations", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.1008/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:56Z", "role": "primary", "sha256": "e8930e90d2d78f0d6ef35acc487ba355d3798cdafbcc1d2637258bad81407d41", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.1008/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.13451", "identifiers": {"arxiv": "2411.13451"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.13451"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.1008/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.1008.pdf"}], "name": "AdaptAgent: Adapting Multimodal Web Agents with Few-Shot Learning from Human Demonstrations", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-11-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adapts multimodal web agents to unfamiliar websites using a small number of human demonstrations through prompting or meta-learning.", "topics": ["web-navigation", "data-generation"], "updated_at": "2024-11-20", "url": "https://arxiv.org/abs/2411.13451"}
{"aliases": [], "authors": ["Davide Paglieri", "Bartłomiej Cupiał", "Samuel Coward", "Ulyana Piterbarg", "Maciej Wolczyk", "Akbir Khan", "Eduardo Pignatelli", "Łukasz Kuciński", "Lerrel Pinto", "Rob Fergus", "Jakob Nicolaus Foerster", "Jack Parker-Holder", "Tim Rocktäschel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b2aa208733edb5cc82d996cd1ace1b05a1d36e6cf3fcb71df377ff835bbba87", "submitted_at": "2024-11-20T18:54:32Z", "updated_at": "2025-04-01T14:45:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.13543v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2411.13543", "identifiers": {"arxiv": "2411.13543"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.13543"}], "name": "BALROG: Benchmarking Agentic LLM and VLM Reasoning On Games", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-01", "url": "https://arxiv.org/abs/2411.13543"}
{"aliases": [], "authors": ["Anthony Nguyen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b639d016daf616a5f58656d2a291505cd82f7765c47638a515e359210fd56779", "submitted_at": "2024-11-18T05:47:12Z", "updated_at": "2025-09-11T16:37:00Z", "version": 7, "versioned_url": "https://arxiv.org/abs/2411.13591v7"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.13591", "identifiers": {"arxiv": "2411.13591"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.13591"}], "name": "Improved GUI Grounding via Iterative Narrowing", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-11-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses iterative visual narrowing to improve localization of requested GUI elements with general or specially trained vision-language models.", "topics": ["grounding"], "updated_at": "2025-09-11", "url": "https://arxiv.org/abs/2411.13591"}
{"aliases": [], "authors": ["Yuke Zhu", "Chi Xie", "Shuang Liang", "Bo Zheng", "Sheng Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c881b3f605f4fe630e31292d9f5c068f1ee3a2bff3bbd2f65d7b9f87f6d76020", "submitted_at": "2024-11-21T15:37:52Z", "updated_at": "2024-11-21T15:37:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.14228v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2411.14228", "identifiers": {"arxiv": "2411.14228"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.14228"}], "name": "FocusLLaVA: A Coarse-to-Fine Approach for Efficient and Effective Visual Token Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-21", "url": "https://arxiv.org/abs/2411.14228"}
{"aliases": [], "authors": ["Junhong Shen", "Atishay Jain", "Zedian Xiao", "Ishan Amlekar", "Mouad Hadji", "Aaron Podolny", "Ameet Talwalkar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ef5e3848c3395a5620c5af422e6ebf198e405cfde1314cfac50d347daf68a429", "submitted_at": "2024-11-22T15:26:23Z", "updated_at": "2024-12-05T02:00:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.15004v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.15004", "identifiers": {"arxiv": "2411.15004"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.15004"}], "name": "ScribeAgent: Towards Specialized Web Agents Using Production-Scale Workflow Data", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-11-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies fine-tuning open language models on production web-workflow traces, including training choices for specialized web agents.", "topics": ["workflow", "web-navigation"], "updated_at": "2024-12-05", "url": "https://arxiv.org/abs/2411.15004"}
{"aliases": [], "authors": ["Leonidas Gee", "Wing Yan Li", "Viktoriia Sharmanska", "Novi Quadrianto"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "072a4444cbe5572db71e5d6482e8f2a4fbb0832ec663a12b848524c92ecedf3f", "submitted_at": "2024-11-23T00:47:13Z", "updated_at": "2025-11-27T14:09:20Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2411.15397v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2411.15397", "identifiers": {"arxiv": "2411.15397"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2411.15397"}], "name": "Visual-Word Tokenizer: Beyond Fixed Sets of Tokens in Vision Transformers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-27", "url": "https://arxiv.org/abs/2411.15397"}
{"aliases": [], "authors": ["Kevin Qinghong Lin", "Linjie Li", "Difei Gao", "Zhengyuan Yang", "Shiwei Wu", "Zechen Bai", "Weixian Lei", "Lijuan Wang", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8c8f5c0a7fc5f194ffa6ca914619b9402c947d15f100b124505883ad8ff23ac7", "submitted_at": "2024-11-26T14:29:47Z", "updated_at": "2024-11-26T14:29:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.17465v1"}, "publication": {"authors": ["Kevin Qinghong Lin", "Linjie Li", "Difei Gao", "Zhengyuan Yang", "Shiwei Wu", "Zechen Bai", "Stan Weixian Lei", "Lijuan Wang", "Mike Zheng Shou"], "date_basis": "Year from publisher citation metadata and month from the same page's embedded BibTeX; no day is supplied.", "date_parts": [2025, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher page directly links this arXiv identifier.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "19498-19508", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2025/papers/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.pdf", "publisher": null, "series": null, "title": "ShowUI: One Vision-Language-Action Model for GUI Visual Agent", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:00Z", "role": "primary", "sha256": "2a79195c42f55ef363988febb2b7cae1d348c4450a443462b75c96fd20c17256", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2411.17465", "identifiers": {"arxiv": "2411.17465"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.17465"}, {"rel": "publication", "url": "https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html"}, {"rel": "publication_pdf", "url": "https://openaccess.thecvf.com/content/CVPR2025/papers/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.pdf"}], "name": "ShowUI: One Vision-Language-Action Model for GUI Visual Agent", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-11-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines visual token selection, interleaved visual-action history, and curated instruction data in a compact GUI action model.", "topics": ["grounding", "efficiency"], "updated_at": "2024-11-26", "url": "https://arxiv.org/abs/2411.17465"}
{"aliases": [], "authors": ["Chaoyun Zhang", "Shilin He", "Jiaxu Qian", "Bowen Li", "Liqun Li", "Si Qin", "Yu Kang", "Minghua Ma", "Guyue Liu", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f692aa06152820ae60359fc2cc52aaed2ca8a97fe1efb9b8037de973956467a", "submitted_at": "2024-11-27T12:13:39Z", "updated_at": "2025-05-06T15:08:00Z", "version": 12, "versioned_url": "https://arxiv.org/abs/2411.18279v12"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2411.18279", "identifiers": {"arxiv": "2411.18279"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.18279"}], "name": "Large Language Model-Brained GUI Agents: A Survey", "notes": "", "published_at": "2024-11-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Survey of LLM-brained GUI agents that carry out natural-language instructions across web, mobile, and desktop interfaces, organizing research and industry progress.", "topics": ["survey"], "updated_at": "2025-05-06", "url": "https://arxiv.org/abs/2411.18279"}
{"aliases": [], "authors": ["Weikai Xie", "Li Zhang", "Shihe Wang", "Rongjie Yi", "Mengwei Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c22d6a9aafb89e5628198fb4f6ea50fe40b1e955a2fa367571459753cea61df", "submitted_at": "2024-11-30T08:55:39Z", "updated_at": "2024-11-30T08:55:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.00402v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "arxiv:2412.00402", "identifiers": {"arxiv": "2412.00402"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.00402"}], "name": "DroidCall: A Dataset for LLM-powered Android Intent Invocation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-11-30", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces data and a generation pipeline for translating natural-language requests into Android intent invocations using compact language models.", "topics": ["tool-integration", "data-generation"], "updated_at": "2024-11-30", "url": "https://arxiv.org/abs/2412.00402"}
{"aliases": [], "authors": ["Yiqin Wang", "Haoji Zhang", "Jingqi Tian", "Yansong Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a4c750017de421d3ca9bf119834dfb96d2fe8a0610f12b30d2fae603e26c66b", "submitted_at": "2024-12-02T08:35:31Z", "updated_at": "2024-12-02T08:35:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.01268v1"}, "publication": {"authors": ["Yiqin Wang", "Haoji Zhang", "Jingqi Tian", "Yansong Tang"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.76", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.76/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "1461-1473", "pdf_url": "https://aclanthology.org/2025.findings-acl.76.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Ponder & Press: Advancing Visual GUI Agent towards General Computer Control", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.76/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["cross-platform", "web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:57Z", "role": "primary", "sha256": "0ec2dd7116411e80f842627d047d06cf60d5921d91350daeb251a9c95a8f7884", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.76/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.01268", "identifiers": {"arxiv": "2412.01268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.01268"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.76/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.76.pdf"}], "name": "Ponder & Press: Advancing Visual GUI Agent towards General Computer Control", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates interpretation of user instructions from precise GUI localization using general-purpose and specialized visual-language components.", "topics": ["grounding", "planning"], "updated_at": "2024-12-02", "url": "https://arxiv.org/abs/2412.01268"}
{"aliases": [], "authors": ["Erick Galinkin", "Martin Sablotny"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "92d55a4970abef7ec4de985473ffd9f28b32b7c8e3271210af7b30f62b696455", "submitted_at": "2024-12-02T14:35:43Z", "updated_at": "2024-12-02T14:35:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.01547v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2412.01547", "identifiers": {"arxiv": "2412.01547"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.01547"}], "name": "Improved Large Language Model Jailbreak Detection via Pretrained Embeddings", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-02", "url": "https://arxiv.org/abs/2412.01547"}
{"aliases": [], "authors": ["Yiyuan Pan", "Yunzhe Xu", "Zhe Liu", "Hesheng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60ef3b209e3106b94c11703c505501e3530df8247c4df5418514490a60b75227", "submitted_at": "2024-11-30T16:49:14Z", "updated_at": "2024-12-25T08:59:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.01857v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.01857", "identifiers": {"arxiv": "2412.01857"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.01857"}], "name": "Planning from Imagination: Episodic Simulation and Episodic Memory for Vision-and-Language Navigation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-25", "url": "https://arxiv.org/abs/2412.01857"}
{"aliases": [], "authors": ["Muhammad Khalifa", "Yi-Chern Tan", "Arash Ahmadian", "Tom Hosking", "Honglak Lee", "Lu Wang", "Ahmet Üstün", "Tom Sherborne", "Matthias Gallé"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f029b027919c39b126876b11161a23859cb6e2952c90e80754050bf9c48d8989", "submitted_at": "2024-12-05T13:12:51Z", "updated_at": "2025-02-03T20:31:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.04144v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.04144", "identifiers": {"arxiv": "2412.04144"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.04144"}], "name": "If You Can't Use Them, Recycle Them: Optimizing Merging at Scale Mitigates Performance Tradeoffs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-03", "url": "https://arxiv.org/abs/2412.04144"}
{"aliases": [], "authors": ["Rong Li", "Shijie Li", "Lingdong Kong", "Xulei Yang", "Junwei Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0c923a3d79900fa02279e24e31da14a1d22d6d971eb7b9be4fa06339238fc262", "submitted_at": "2024-12-05T17:58:43Z", "updated_at": "2025-05-29T14:14:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.04383v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.04383", "identifiers": {"arxiv": "2412.04383"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.04383"}], "name": "SeeGround: See and Ground for Zero-Shot Open-Vocabulary 3D Visual Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-29", "url": "https://arxiv.org/abs/2412.04383"}
{"aliases": [], "authors": ["Yiheng Xu", "Zekun Wang", "Junli Wang", "Dunjie Lu", "Tianbao Xie", "Amrita Saha", "Doyen Sahoo", "Tao Yu", "Caiming Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89dbdf874ace87db12fa321e67ecf818d3964622b045ea7d95e109101820d407", "submitted_at": "2024-12-05T18:58:26Z", "updated_at": "2025-05-05T16:17:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.04454v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.04454", "identifiers": {"arxiv": "2412.04454"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.04454"}], "name": "Aguvis: Unified Pure Vision Agents for Autonomous GUI Interaction", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2024-12-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses screenshots, a unified action space, and staged grounding and reasoning training for cross-platform GUI interaction.", "topics": ["grounding", "planning", "data-generation"], "updated_at": "2025-05-05", "url": "https://arxiv.org/abs/2412.04454"}
{"aliases": [], "authors": ["Juan Rodriguez", "Xiangru Jian", "Siba Smarak Panigrahi", "Tianyu Zhang", "Aarash Feizi", "Abhay Puri", "Akshay Kalkunte", "François Savard", "Ahmed Masry", "Shravan Nayak", "Rabiul Awal", "Mahsa Massoud", "Amirhossein Abaskohi", "Zichao Li", "Suyuchen Wang", "Pierre-André Noël", "Mats Leon Richter", "Saverio Vadacchino", "Shubham Agarwal", "Sanket Biswas", "Sara Shanian", "Ying Zhang", "Noah Bolger", "Kurt MacDonald", "Simon Fauvel", "Sathwik Tejaswi", "Srinivas Sunkara", "Joao Monteiro", "Krishnamurthy DJ Dvijotham", "Torsten Scholak", "Nicolas Chapados", "Sepideh Kharagani", "Sean Hughes", "M. Özsu", "Siva Reddy", "Marco Pedersoli", "Yoshua Bengio", "Christopher Pal", "Issam Laradji", "Spandana Gella", "Perouz Taslakian", "David Vazquez", "Sai Rajeswar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "84b72716d09298ff04a3b54e4c17f56b86a54927fddcb8f481890dbea7fb7e2b", "submitted_at": "2024-12-05T21:41:20Z", "updated_at": "2025-03-17T16:32:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.04626v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.04626", "identifiers": {"arxiv": "2412.04626"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.04626"}], "name": "BigDocs: An Open Dataset for Training Multimodal Models on Document and Code Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-17", "url": "https://arxiv.org/abs/2412.04626"}
{"aliases": [], "authors": ["Nicholas J. Boyer", "Christopher Shepard", "Ruiyi Zhou", "Jianhang Xu", "Yosuke Kanai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b60f486694c7a9c77fc27ca1b5763d2af6f84084f5ad4dfa8066c94eb2ffc762", "submitted_at": "2024-12-06T18:49:53Z", "updated_at": "2024-12-06T18:49:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.05260v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.05260", "identifiers": {"arxiv": "2412.05260"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.05260"}], "name": "Machine-Learning Electron Dynamics with Moment Propagation Theory: Application to Optical Absorption Spectrum Computation using Real-Time TDDFT", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-06", "url": "https://arxiv.org/abs/2412.05260"}
{"aliases": [], "authors": ["Thibault Le Sellier De Chezelles", "Maxime Gasse", "Alexandre Drouin", "Massimo Caccia", "Léo Boisvert", "Megh Thakkar", "Tom Marty", "Rim Assouel", "Sahar Omidi Shayegan", "Lawrence Keunho Jang", "Xing Han Lù", "Ori Yoran", "Dehan Kong", "Frank F. Xu", "Siva Reddy", "Quentin Cappart", "Graham Neubig", "Ruslan Salakhutdinov", "Nicolas Chapados", "Alexandre Lacoste"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "572a993cd37a83bf7100dca71f0a44473ed6547f521409abf0f4ee22733a28cc", "submitted_at": "2024-12-06T23:43:59Z", "updated_at": "2025-02-28T16:02:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2412.05467v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.05467", "identifiers": {"arxiv": "2412.05467"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.05467"}], "name": "The BrowserGym Ecosystem for Web Agent Research", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes BrowserGym's shared web interaction environment and AgentLab tooling for consistent experiments across multiple agent benchmarks.", "topics": ["evaluation", "sandbox"], "updated_at": "2025-02-28", "url": "https://arxiv.org/abs/2412.05467"}
{"aliases": [], "authors": ["Ao Wang", "Fengyuan Sun", "Hui Chen", "Zijia Lin", "Jungong Han", "Guiguang Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f3be4ee4be9a21ff80e0ba1f679c7318d75cd33bbcf7d3a9f316ff6aec90f93", "submitted_at": "2024-12-08T05:29:39Z", "updated_at": "2024-12-08T05:29:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.05819v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.05819", "identifiers": {"arxiv": "2412.05819"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.05819"}], "name": "[CLS] Token Tells Everything Needed for Training-free Efficient MLLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-08", "url": "https://arxiv.org/abs/2412.05819"}
{"aliases": [], "authors": ["Jiazhao Zhang", "Kunyu Wang", "Shaoan Wang", "Minghan Li", "Haoran Liu", "Songlin Wei", "Zhongyuan Wang", "Zhizheng Zhang", "He Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db7005e67ec26142c94fc915d0410d211e1a2ae75277c93a893c1206ed84e3cc", "submitted_at": "2024-12-09T05:55:55Z", "updated_at": "2025-02-06T10:14:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.06224v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.06224", "identifiers": {"arxiv": "2412.06224"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.06224"}], "name": "Uni-NaVid: A Video-based Vision-Language-Action Model for Unifying Embodied Navigation Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-06", "url": "https://arxiv.org/abs/2412.06224"}
{"aliases": [], "authors": ["Meera Hahn", "Wenjun Zeng", "Nithish Kannen", "Rich Galt", "Kartikeya Badola", "Been Kim", "Zi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "045ce24c86866883018771f2ff9343312cb415385e586d218bf321f057237a1a", "submitted_at": "2024-12-09T18:56:32Z", "updated_at": "2025-10-24T02:32:07Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.06771v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.06771", "identifiers": {"arxiv": "2412.06771"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.06771"}], "name": "Proactive Agents for Multi-Turn Text-to-Image Generation Under Uncertainty", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-24", "url": "https://arxiv.org/abs/2412.06771"}
{"aliases": [], "authors": ["Christian Duffee", "Jordan Athas", "Yixin Shao", "Noraica Davila Melendez", "Eleonora Raimondo", "Jordan A. Katine", "Kerem Y. Camsari", "Giovanni Finocchio", "Pedram Khalili Amiri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "59fe8531ff7f37261be8018c26aaf439a0dbe5b5b22b6ec08a9c250d664ad2e9", "submitted_at": "2024-12-11T01:55:13Z", "updated_at": "2024-12-11T01:55:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.08017v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.08017", "identifiers": {"arxiv": "2412.08017"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.08017"}], "name": "Integrated probabilistic computer using voltage-controlled magnetic tunnel junctions as its entropy source", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-11", "url": "https://arxiv.org/abs/2412.08017"}
{"aliases": [], "authors": ["Stephen C. Wein", "Timothée Goubault de Brugière", "Luka Music", "Pascale Senellart", "Boris Bourdoncle", "Shane Mansfield"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "38cb8ddd16d03b9fc6d9e2168198b1a9cfd97ab253aa66f6e74216e339876d20", "submitted_at": "2024-12-11T18:34:53Z", "updated_at": "2025-12-19T16:12:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.08611v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.08611", "identifiers": {"arxiv": "2412.08611"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.08611"}], "name": "Minimizing resource overhead in fusion-based quantum computation using hybrid spin-photon devices", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-19", "url": "https://arxiv.org/abs/2412.08611"}
{"aliases": [], "authors": ["Huawen Shen", "Chang Liu", "Gengluo Li", "Xinlong Wang", "Yu Zhou", "Can Ma", "Xiangyang Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ee5840bbd1404674f0e9509b305e995c4dece9257e3f2a79348dde9aee99928b", "submitted_at": "2024-12-12T15:29:36Z", "updated_at": "2024-12-12T15:29:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.09362v1"}}, "environments": ["web", "android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.09362", "identifiers": {"arxiv": "2412.09362"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.09362"}], "name": "Falcon-UI: Understanding GUI Before Following User Instructions", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Pretrains GUI-context understanding with instruction-free web interactions before fine-tuning Falcon-UI on instructed GUI tasks.", "topics": ["screen-understanding", "data-generation"], "updated_at": "2024-12-12", "url": "https://arxiv.org/abs/2412.09362"}
{"aliases": [], "authors": ["Han Wang", "Yuxiang Nie", "Yongjie Ye", "Deng GuanYu", "Yanjie Wang", "Shuai Li", "Haiyang Yu", "Jinghui Lu", "Can Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe3e1d2d2f60befb96f47bfc8b0d8d372388a2cca8bd8ab8e90efa00cae714c9", "submitted_at": "2024-12-12T18:20:41Z", "updated_at": "2024-12-12T18:20:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.09530v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.09530", "identifiers": {"arxiv": "2412.09530"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.09530"}], "name": "Dynamic-VLM: Simple Dynamic Visual Token Compression for VideoLLM", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-12", "url": "https://arxiv.org/abs/2412.09530"}
{"aliases": [], "authors": ["Yiheng Xu", "Dunjie Lu", "Zhennan Shen", "Junli Wang", "Zekun Wang", "Yuchen Mao", "Caiming Xiong", "Tao Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ee630d434cef63f63c368752aeae14b0a60d49c3fde33b3a79ecbfc8fc4d06e", "submitted_at": "2024-12-12T18:59:27Z", "updated_at": "2025-03-03T18:59:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.09605v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.09605", "identifiers": {"arxiv": "2412.09605"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.09605"}], "name": "AgentTrek: Agent Trajectory Synthesis via Guiding Replay with Web Tutorials", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Converts public tutorials into task specifications and executes them with an agent to synthesize multimodal interaction trajectories.", "topics": ["data-generation", "web-navigation"], "updated_at": "2025-03-03", "url": "https://arxiv.org/abs/2412.09605"}
{"aliases": [], "authors": ["Chenyu Yang", "Xuan Dong", "Xizhou Zhu", "Weijie Su", "Jiahao Wang", "Hao Tian", "Zhe Chen", "Wenhai Wang", "Lewei Lu", "Jifeng Dai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2012d9ca5d9f121825596f35e63131cdb69b5db466d1b8b1a06f5e405b5e1601", "submitted_at": "2024-12-12T18:59:40Z", "updated_at": "2024-12-12T18:59:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.09613v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.09613", "identifiers": {"arxiv": "2412.09613"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.09613"}], "name": "PVC: Progressive Visual Token Compression for Unified Image and Video Processing in Large Vision-Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-12", "url": "https://arxiv.org/abs/2412.09613"}
{"aliases": [], "authors": ["Lu Wang", "Fangkai Yang", "Chaoyun Zhang", "Junting Lu", "Jiaxu Qian", "Shilin He", "Pu Zhao", "Bo Qiao", "Ray Huang", "Si Qin", "Qisheng Su", "Jiayi Ye", "Yudi Zhang", "Jian-Guang Lou", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54051d974e388ed2a26f1421b8c1fc79472bb941810519039ef48e90bf4761b6", "submitted_at": "2024-12-13T11:19:56Z", "updated_at": "2025-01-13T06:47:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.10047v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:28Z", "role": "primary", "sha256": "aaddc3cae2fcadc9780390fceb663c0200ecb042639e2c69bccea19607cd78ea", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2412.10047v2"}], "id": "arxiv:2412.10047", "identifiers": {"arxiv": "2412.10047"}, "kind": "paper", "last_verified_at": "2026-10-17T09:13:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10047"}], "name": "Large Action Models: From Inception to Implementation", "notes": "Reviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2024-12-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses a Windows-agent case study to describe a development workflow for large action models, covering data collection, training, grounding, environment integration, and evaluation.", "topics": ["automation", "data-generation", "evaluation"], "updated_at": "2025-01-13", "url": "https://arxiv.org/abs/2412.10047"}
{"aliases": [], "authors": ["Zhiqi Ge", "Juncheng Li", "Xinglei Pang", "Minghe Gao", "Kaihang Pan", "Wang Lin", "Hao Fei", "Wenqiao Zhang", "Siliang Tang", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e7cab57b118ab76a9026e7ee800709f4e93969e19ad18f549f8b75ee657622b", "submitted_at": "2024-12-13T18:40:10Z", "updated_at": "2025-02-03T15:23:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.10342v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.10342", "identifiers": {"arxiv": "2412.10342"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10342"}], "name": "Iris: Breaking GUI Complexity with Adaptive Focus and Self-Refining", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines information-sensitive image cropping with a learning loop between element description and grounding for complex interface understanding.", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2025-02-03", "url": "https://arxiv.org/abs/2412.10342"}
{"aliases": [], "authors": ["Gagan Bansal", "Jennifer Wortman Vaughan", "Saleema Amershi", "Eric Horvitz", "Adam Fourney", "Hussein Mozannar", "Victor Dibia", "Daniel S. Weld"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e0db5517a22b8d33d2082226b4062cb748c1b8871cf22f3f81a890c38e44626", "submitted_at": "2024-11-28T01:21:26Z", "updated_at": "2024-11-28T01:21:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.10380v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.10380", "identifiers": {"arxiv": "2412.10380"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.10380"}], "name": "Challenges in Human-Agent Communication", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-11-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-11-28", "url": "https://arxiv.org/abs/2412.10380"}
{"aliases": [], "authors": ["Chi Zhang", "Jiajun Song", "Siyu Li", "Yitao Liang", "Yuxi Ma", "Wei Wang", "Yixin Zhu", "Song-Chun Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9335a1eda851583681ce4c9dc590e0795247d510e3f6b5db41beee0e36321bf5", "submitted_at": "2024-12-14T04:20:47Z", "updated_at": "2024-12-14T04:20:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.10673v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.10673", "identifiers": {"arxiv": "2412.10673"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.10673"}], "name": "Proposing and solving olympiad geometry with guided tree search", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-14", "url": "https://arxiv.org/abs/2412.10673"}
{"aliases": [], "authors": ["Jiarun Liu", "Jia Hao", "Chunhong Zhang", "Zheng Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e2abbae042d20381aa218e6a2b14a1fc28074543e430ad09031237d22587be7a", "submitted_at": "2024-12-14T08:25:28Z", "updated_at": "2024-12-14T08:25:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.10742v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.10742", "identifiers": {"arxiv": "2412.10742"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10742"}], "name": "WEPO: Web Element Preference Optimization for LLM-based Web Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses web-element preference optimization with sampled negative HTML elements to align high-level user intent with navigation actions.", "topics": ["grounding", "web-navigation"], "updated_at": "2024-12-14", "url": "https://arxiv.org/abs/2412.10742"}
{"aliases": [], "authors": ["Hai-Ming Xu", "Qi Chen", "Lei Wang", "Lingqiao Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6e0aefcd3eca9e37d25a2c5195c0d502caec9be83b8f6b4b7ee5778bff9e150", "submitted_at": "2024-12-14T14:30:05Z", "updated_at": "2024-12-14T14:30:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.10840v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.10840", "identifiers": {"arxiv": "2412.10840"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10840"}], "name": "Attention-driven GUI Grounding: Leveraging Pretrained Multimodal Large Language Models without Fine-Tuning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Locates GUI targets by identifying and aggregating attention maps in pretrained multimodal models without task-specific fine-tuning.", "topics": ["grounding"], "updated_at": "2024-12-14", "url": "https://arxiv.org/abs/2412.10840"}
{"aliases": [], "authors": ["Jiya Manchanda", "Laura Boettcher", "Matheus Westphalen", "Jasser Jasser"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ad756bc4e799f251a178edabb1e59e1405a25d9369d442de5fccaa0dee501177", "submitted_at": "2024-12-16T17:32:11Z", "updated_at": "2025-10-14T01:48:35Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.12004v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.12004", "identifiers": {"arxiv": "2412.12004"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.12004"}], "name": "The Open Source Advantage in Large Language Models (LLMs)", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-14", "url": "https://arxiv.org/abs/2412.12004"}
{"aliases": [], "authors": ["Tiannan Wang", "Meiling Tao", "Ruoyu Fang", "Huilin Wang", "Shuai Wang", "Yuchen Eleanor Jiang", "Wangchunshu Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6187ac36c2c4e7f78155cc6a1aea504aeb7350386697031220195d1aeaba9ea", "submitted_at": "2024-12-17T17:17:03Z", "updated_at": "2024-12-17T17:17:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.13103v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.13103", "identifiers": {"arxiv": "2412.13103"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.13103"}], "name": "AI PERSONA: Towards Life-long Personalization of LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-17", "url": "https://arxiv.org/abs/2412.13103"}
{"aliases": [], "authors": ["Sheng Yin", "Xianghe Pang", "Yuanzhuo Ding", "Menglan Chen", "Yutong Bi", "Yichen Xiong", "Wenhao Huang", "Zhen Xiang", "Jing Shao", "Siheng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ad79285e547b3229c91db604c40634947fa9558eb594a3105efc75436e2369b", "submitted_at": "2024-12-17T18:55:58Z", "updated_at": "2025-10-31T08:18:50Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2412.13178v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.13178", "identifiers": {"arxiv": "2412.13178"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.13178"}], "name": "SafeAgentBench: A Benchmark for Safe Task Planning of Embodied LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-31", "url": "https://arxiv.org/abs/2412.13178"}
{"aliases": [], "authors": ["Yifei Zhou", "Qianlan Yang", "Kaixiang Lin", "Min Bai", "Xiong Zhou", "Yu-Xiong Wang", "Sergey Levine", "Erran Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "362d2a470603bece43d2205e56dd633e7310cfad2ba2d91b0b765abcbcc41b2f", "submitted_at": "2024-12-17T18:59:50Z", "updated_at": "2024-12-17T18:59:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.13194v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.13194", "identifiers": {"arxiv": "2412.13194"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.13194"}], "name": "Proposer-Agent-Evaluator(PAE): Autonomous Skill Discovery For Foundation Model Internet Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines autonomous task proposal, web interaction, and visual success evaluation to discover skills and improve browser-agent policies.", "topics": ["data-generation", "reinforcement-learning", "web-navigation"], "updated_at": "2024-12-17", "url": "https://arxiv.org/abs/2412.13194"}
{"aliases": [], "authors": ["Dang Nguyen", "Jian Chen", "Yu Wang", "Gang Wu", "Namyong Park", "Zhengmian Hu", "Hanjia Lyu", "Junda Wu", "Ryan Aponte", "Yu Xia", "Xintong Li", "Jing Shi", "Hongjie Chen", "Viet Dac Lai", "Zhouhang Xie", "Sungchul Kim", "Ruiyi Zhang", "Tong Yu", "Mehrab Tanjim", "Nesreen K. Ahmed", "Puneet Mathur", "Seunghyun Yoon", "Lina Yao", "Branislav Kveton", "Jihyung Kil", "Thien Huu Nguyen", "Trung Bui", "Tianyi Zhou", "Ryan A. Rossi", "Franck Dernoncourt"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae8f83d0490e70db1ae9c55c071935fb1f59e1358335c1bd3f50835611f4360c", "submitted_at": "2024-12-18T04:48:28Z", "updated_at": "2025-09-26T21:01:59Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.13501v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.13501", "identifiers": {"arxiv": "2412.13501"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.13501"}], "name": "GUI Agents: A Survey", "notes": "", "published_at": "2024-12-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Categorizes GUI-agent benchmarks, evaluation metrics, architectures, and training methods, and proposes a unified perception, reasoning, planning, and acting framework.", "topics": ["survey"], "updated_at": "2025-09-26", "url": "https://arxiv.org/abs/2412.13501"}
{"aliases": [], "authors": ["Dimitrios Mallis", "Ahmet Serdar Karadeniz", "Sebastian Cavada", "Danila Rukhovich", "Niki Foteinopoulou", "Kseniya Cherenkova", "Anis Kacem", "Djamila Aouada"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f60272fa5cd2d012e0203f53440d834e996011e4577ed087e4783eae189a9c83", "submitted_at": "2024-12-18T12:57:56Z", "updated_at": "2025-08-26T14:18:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.13810v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.13810", "identifiers": {"arxiv": "2412.13810"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.13810"}], "name": "CAD-Assistant: Tool-Augmented VLLMs as Generic CAD Task Solvers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-26", "url": "https://arxiv.org/abs/2412.13810"}
{"aliases": [], "authors": ["Frank F. Xu", "Yufan Song", "Boxuan Li", "Yuxuan Tang", "Kritanjali Jain", "Mengxue Bao", "Zora Z. Wang", "Xuhui Zhou", "Zhitong Guo", "Murong Cao", "Mingyang Yang", "Hao Yang Lu", "Amaad Martin", "Zhe Su", "Leander Maben", "Raj Mehta", "Wayne Chi", "Lawrence Jang", "Yiqing Xie", "Shuyan Zhou", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77f91dd9ec9f50fb5526a78cfe270a7a1705d5bd83a2aef26644453740c005d5", "submitted_at": "2024-12-18T18:55:40Z", "updated_at": "2025-09-10T08:35:19Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.14161v3"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.14161", "identifiers": {"arxiv": "2412.14161"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.14161"}], "name": "TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Simulates a software-company workplace to evaluate agents on professional tasks involving websites, programs, files, and coworker communication.", "topics": ["evaluation", "long-horizon", "workflow"], "updated_at": "2025-09-10", "url": "https://arxiv.org/abs/2412.14161"}
{"aliases": [], "authors": ["Yuhao Du", "Shunian Chen", "Wenbo Zan", "Peizhao Li", "Mingxuan Wang", "Dingjie Song", "Bo Li", "Yan Hu", "Benyou Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a15e603ca3c91e31389fad9188f25a78de9496d92ef9e1583500cd29794ed58d", "submitted_at": "2024-12-16T14:34:02Z", "updated_at": "2024-12-16T14:34:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.14203v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.14203", "identifiers": {"arxiv": "2412.14203"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.14203"}], "name": "BlenderLLM: Training Large Language Models for Computer-Aided Design with Self-improvement", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-16", "url": "https://arxiv.org/abs/2412.14203"}
{"aliases": [], "authors": ["Maojun Sun", "Ruijian Han", "Binyan Jiang", "Houduo Qi", "Defeng Sun", "Yancheng Yuan", "Jian Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c447cc61f61c098a817fa917ba10fd22864e971bd59b7bf0431f89885e34001", "submitted_at": "2024-12-18T15:03:26Z", "updated_at": "2025-09-14T04:25:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.14222v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.14222", "identifiers": {"arxiv": "2412.14222"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.14222"}], "name": "A Survey on Large Language Model-based Agents for Statistics and Data Science", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-14", "url": "https://arxiv.org/abs/2412.14222"}
{"aliases": [], "authors": ["Zhexin Zhang", "Shiyao Cui", "Yida Lu", "Jingzhuo Zhou", "Junxiao Yang", "Hongning Wang", "Minlie Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4a4ecbf49d71bdb8ca15b367627e739d3b342a64f90834f9dd5d0d7b65817dc3", "submitted_at": "2024-12-19T02:35:15Z", "updated_at": "2025-05-20T05:58:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.14470v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.14470", "identifiers": {"arxiv": "2412.14470"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.14470"}], "name": "Agent-SafetyBench: Evaluating the Safety of LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-20", "url": "https://arxiv.org/abs/2412.14470"}
{"aliases": [], "authors": ["Yuhao Yang", "Yue Wang", "Dongxu Li", "Ziyang Luo", "Bei Chen", "Chao Huang", "Junnan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b49d45aa539c224846a7e14cee9b537d032228e19e86a9cb3cdd9f5004aa429", "submitted_at": "2024-12-20T07:16:57Z", "updated_at": "2025-07-08T08:49:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.16256v2"}, "publication": {"authors": ["Yuhao Yang", "Yue Wang", "Dongxu Li", "Ziyang Luo", "Bei Chen", "Chao Huang", "Junnan Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.1152", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.1152/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "22418-22433", "pdf_url": "https://aclanthology.org/2025.findings-acl.1152.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Aria-UI: Visual Grounding for GUI Instructions", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.1152/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:57Z", "role": "primary", "sha256": "e24a63769d700caa2d27341dd2634b990cdbc53c7cf557817b3679dc63faef6f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.1152/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.16256", "identifiers": {"arxiv": "2412.16256"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.16256"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.1152/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.1152.pdf"}], "name": "Aria-UI: Visual Grounding for GUI Instructions", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a vision-based grounding model trained with diverse synthetic instructions and action histories for context-sensitive GUI targeting.", "topics": ["grounding", "memory"], "updated_at": "2025-07-08", "url": "https://arxiv.org/abs/2412.16256"}
{"aliases": [], "authors": ["Nilanjana Das", "Edward Raff", "Aman Chadha", "Manas Gaur"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8c96e32d280d7b12add520ade56e4ecc1cb37c4fc84018693c572c7b3ab1fa9c", "submitted_at": "2024-12-20T21:43:52Z", "updated_at": "2025-05-29T05:54:54Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.16359v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.16359", "identifiers": {"arxiv": "2412.16359"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.16359"}], "name": "Human-Readable Adversarial Prompts: An Investigation into LLM Vulnerabilities Using Situational Context", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-29", "url": "https://arxiv.org/abs/2412.16359"}
{"aliases": [], "authors": ["Yanheng He", "Jiahe Jin", "Shijie Xia", "Jiadi Su", "Runze Fan", "Haoyang Zou", "Xiangkun Hu", "Pengfei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db567d35887fe5367e456b35bc347400b1bc2d147730e9d34c158e1fd7c8dd6c", "submitted_at": "2024-12-23T14:02:12Z", "updated_at": "2024-12-23T14:02:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.17589v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.17589", "identifiers": {"arxiv": "2412.17589"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.17589"}], "name": "PC Agent: While You Sleep, AI Works -- A Cognitive Journey into Digital World", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines PC Tracker interaction capture, completion of action and reasoning annotations, and planning and grounding agents for desktop workflows.", "topics": ["recording", "data-generation", "planning"], "updated_at": "2024-12-23", "url": "https://arxiv.org/abs/2412.17589"}
{"aliases": [], "authors": ["Hao Wen", "Shizuo Tian", "Borislav Pavlov", "Wenjie Du", "Yixuan Li", "Ge Chang", "Shanhui Zhao", "Jiacheng Liu", "Yunxin Liu", "Ya-Qin Zhang", "Yuanchun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3190c452dbfa24b131cc34f66192bbdb9c538d30a1919082a9fed334f6169fff", "submitted_at": "2024-12-24T02:54:56Z", "updated_at": "2025-05-06T12:24:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.18116v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.18116", "identifiers": {"arxiv": "2412.18116"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.18116"}], "name": "AutoDroid-V2: Boosting SLM-based GUI Agents via Code Generation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Recasts mobile UI automation as on-device code generation, using app-specific documentation and synthetic tasks to guide smaller models.", "topics": ["mobile-navigation", "tool-integration"], "updated_at": "2025-05-06", "url": "https://arxiv.org/abs/2412.18116"}
{"aliases": [], "authors": ["Kangjia Zhao", "Jiahui Song", "Leigang Sha", "Haozhan Shen", "Zhi Chen", "Tiancheng Zhao", "Xiubo Liang", "Jianwei Yin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e509f015068bcf1294a97236ff3bbde5e3ceb167f3cb032fcfd84fc4a5b01d0", "submitted_at": "2024-12-24T13:41:47Z", "updated_at": "2024-12-24T13:41:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.18426v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2412.18426", "identifiers": {"arxiv": "2412.18426"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.18426"}], "name": "GUI Testing Arena: A Unified Benchmark for Advancing Autonomous GUI Testing Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates GUI testing through test-intent generation, execution, and defect detection using real, modified, and synthetic mobile application data.", "topics": ["evaluation", "automation", "mobile-navigation"], "updated_at": "2024-12-24", "url": "https://arxiv.org/abs/2412.18426"}
{"aliases": [], "authors": ["Wang Qun", "Liu Yang", "Lin Qingquan", "Qu Zhijiu", "Jiang Ling"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d56bef2e3502cb315dab4a64b57c0209223c8abd008a621461dd410b3ae425a", "submitted_at": "2024-12-27T13:32:10Z", "updated_at": "2024-12-27T13:32:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.19638v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.19638", "identifiers": {"arxiv": "2412.19638"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.19638"}], "name": "Xmodel-2 Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-27", "url": "https://arxiv.org/abs/2412.19638"}
{"aliases": [], "authors": ["Qiushi Sun", "Kanzhi Cheng", "Zichen Ding", "Chuanyang Jin", "Yian Wang", "Fangzhi Xu", "Zhenyu Wu", "Chengyou Jia", "Liheng Chen", "Zhoumianze Liu", "Ben Kao", "Guohao Li", "Junxian He", "Yu Qiao", "Zhiyong Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40b5e3090c25f80d8d786f623812459b329acc2772f945419a585ffb27c54db3", "submitted_at": "2024-12-27T16:21:58Z", "updated_at": "2025-06-27T08:25:48Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.19723v3"}, "publication": {"authors": ["Qiushi Sun", "Kanzhi Cheng", "Zichen Ding", "Chuanyang Jin", "Yian Wang", "Fangzhi Xu", "Zhenyu Wu", "Chengyou Jia", "Liheng Chen", "Zhoumianze Liu", "Ben Kao", "Guohao Li", "Junxian He", "Yu Qiao", "Zhiyong Wu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.277", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.277/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "5555-5579", "pdf_url": "https://aclanthology.org/2025.acl-long.277.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "OS-Genesis: Automating GUI Agent Trajectory Construction via Reverse Task Synthesis", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.277/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:57Z", "role": "primary", "sha256": "ea538c35ce448bce6e03932b4a3e20815503c461907234b3107b3f8ccbc5481d", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.277/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.19723", "identifiers": {"arxiv": "2412.19723"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.19723"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.277/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.277.pdf"}], "name": "OS-Genesis: Automating GUI Agent Trajectory Construction via Reverse Task Synthesis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2024-12-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds GUI trajectories by exploring environments first and deriving tasks retrospectively, with reward-model filtering of generated trajectories.", "topics": ["data-generation"], "updated_at": "2025-06-27", "url": "https://arxiv.org/abs/2412.19723"}
{"aliases": [], "authors": ["Linhui Xiao", "Xiaoshan Yang", "Xiangyuan Lan", "Yaowei Wang", "Changsheng Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dac6474da5f9cba43fd05d0860857d07f7f312f979f6dfdd8b9df31bc816d7a9", "submitted_at": "2024-12-28T16:34:35Z", "updated_at": "2026-08-04T09:37:47Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2412.20206v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2412.20206", "identifiers": {"arxiv": "2412.20206"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.20206"}], "name": "Toward Visual Grounding: A Survey", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2412.20206"}
{"aliases": [], "authors": ["Gautam Kumar", "Ashwini Ratnoo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "470bbc3516973a331166d9b5ded35d55699708b539ea2fe81a12f47737c65a6d", "submitted_at": "2024-12-30T05:14:36Z", "updated_at": "2024-12-30T05:14:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.20717v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2412.20717", "identifiers": {"arxiv": "2412.20717"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2412.20717"}], "name": "Closing Speed Computation using Stereo Camera and Applications in Unsignalized T-Intersection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2024-12-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2024-12-30", "url": "https://arxiv.org/abs/2412.20717"}
{"aliases": [], "authors": ["Yuxiang Chai", "Shunye Tang", "Han Xiao", "Weifeng Lin", "Hanhao Li", "Jiayu Zhang", "Liang Liu", "Pengxiang Zhao", "Guangyi Liu", "Guozhi Wang", "Shuai Ren", "Rongduo Han", "Haining Zhang", "Siyuan Huang", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ac7100a0f3cc3f004672eb6923b1f0ecb43c3d98da6959f0ff7514d0fb2f442", "submitted_at": "2025-01-02T09:03:56Z", "updated_at": "2026-01-12T07:01:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.01149v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2501.01149", "identifiers": {"arxiv": "2501.01149"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.01149"}], "name": "A3: Android Agent Arena for Mobile GUI Agents with Essential-State Procedural Evaluation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-01-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Android Agent Arena with online application tasks and procedural evaluation based on reaching essential intermediate states.", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2501.01149"}
{"aliases": [], "authors": ["Dayuan Fu", "Keqing He", "Yejie Wang", "Wentao Hong", "Zhuoma Gongque", "Weihao Zeng", "Wei Wang", "Jingang Wang", "Xunliang Cai", "Weiran Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45d2725c84686673c06e76083f1a145b5ad9a74d3d7e34df016b19071c94fc9d", "submitted_at": "2025-01-03T08:55:19Z", "updated_at": "2025-02-24T12:42:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.01702v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2501.01702", "identifiers": {"arxiv": "2501.01702"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.01702"}], "name": "AgentRefine: Enhancing Agent Generalization through Refinement Tuning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-24", "url": "https://arxiv.org/abs/2501.01702"}
{"aliases": [], "authors": ["Shaolei Zhang", "Qingkai Fang", "Zhe Yang", "Yang Feng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40a28b429af78ed0bda8bb61d0604f4befe061bbb65ecfea0d988c59a0b36821", "submitted_at": "2025-01-07T16:03:14Z", "updated_at": "2025-03-02T15:55:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.03895v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2501.03895", "identifiers": {"arxiv": "2501.03895"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.03895"}], "name": "LLaVA-Mini: Efficient Image and Video Large Multimodal Models with One Vision Token", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-02", "url": "https://arxiv.org/abs/2501.03895"}
{"aliases": [], "authors": ["Yuhang Liu", "Pengxiang Li", "Zishu Wei", "Congkai Xie", "Xueyu Hu", "Xinchen Xu", "Shengyu Zhang", "Xiaotian Han", "Hongxia Yang", "Fei Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "18d5b2e34f18c24e5d60d5564c9b08eced39c1b17bf5dac8d997640786688fba", "submitted_at": "2025-01-08T15:45:21Z", "updated_at": "2025-01-08T15:45:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.04575v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2501.04575", "identifiers": {"arxiv": "2501.04575"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.04575"}], "name": "InfiGUIAgent: A Multimodal Generalist GUI Agent with Native Reasoning and Reflection", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-01-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses staged supervised training to combine GUI understanding and grounding with hierarchical reasoning and expectation-based reflection.", "topics": ["grounding", "planning"], "updated_at": "2025-01-08", "url": "https://arxiv.org/abs/2501.04575"}
{"aliases": [], "authors": ["Xuyang Liu", "Ziming Wang", "Junjie Chen", "Yuhang Han", "Yingyao Wang", "Jiale Yuan", "Jun Song", "Siteng Huang", "Honggang Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "336f1ad4c5ef2824cdd1da5b75d8a394586e6fb9aca0195eb813a8a93a9fde04", "submitted_at": "2025-01-09T11:57:58Z", "updated_at": "2026-01-13T07:37:47Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2501.05179v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}], "id": "arxiv:2501.05179", "identifiers": {"arxiv": "2501.05179"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.05179"}], "name": "Global Compression Commander: Plug-and-Play Inference Acceleration for High-Resolution Large Vision-Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-13", "url": "https://arxiv.org/abs/2501.05179"}
{"aliases": [], "authors": ["Xiaoxi Li", "Guanting Dong", "Jiajie Jin", "Yuyao Zhang", "Yujia Zhou", "Yutao Zhu", "Peitian Zhang", "Zhicheng Dou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9a5c1bfacfb1a1e4c536f19c2aacbbd922fb4ce78816bab626a065505bcaebfc", "submitted_at": "2025-01-09T16:48:17Z", "updated_at": "2025-01-09T16:48:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.05366v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2501.05366", "identifiers": {"arxiv": "2501.05366"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.05366"}], "name": "Search-o1: Agentic Search-Enhanced Large Reasoning Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-09", "url": "https://arxiv.org/abs/2501.05366"}
{"aliases": [], "authors": ["Tushar Aggarwal", "Aarohi Bhand"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6024567fca43c3280ddfa1fe0a2c99e2614a59af6cb9779a3f6c633ff361ad71", "submitted_at": "2025-01-11T10:22:04Z", "updated_at": "2025-01-15T20:43:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.06497v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2501.06497", "identifiers": {"arxiv": "2501.06497"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.06497"}], "name": "PASS: Presentation Automation for Slide Generation and Speech", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-15", "url": "https://arxiv.org/abs/2501.06497"}
{"aliases": [], "authors": ["Juntao Ren", "Priya Sundaresan", "Dorsa Sadigh", "Sanjiban Choudhury", "Jeannette Bohg"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b2f3f499bb204e61f1984f754643b8aea48256ac2a708c474de65386484ee640", "submitted_at": "2025-01-13T01:01:44Z", "updated_at": "2025-10-10T23:31:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.06994v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2501.06994", "identifiers": {"arxiv": "2501.06994"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.06994"}], "name": "Motion Tracks: A Unified Representation for Human-Robot Transfer in Few-Shot Imitation Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-10", "url": "https://arxiv.org/abs/2501.06994"}
{"aliases": [], "authors": ["Jialong Wu", "Wenbiao Yin", "Yong Jiang", "Zhenglin Wang", "Zekun Xi", "Runnan Fang", "Linhai Zhang", "Yulan He", "Deyu Zhou", "Pengjun Xie", "Fei Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "32cf550b1e1d5b671b3a4c834787cb67c32c01f508494c5610dac91e95bd92ba", "submitted_at": "2025-01-13T18:58:07Z", "updated_at": "2025-08-10T05:59:20Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.07572v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2501.07572", "identifiers": {"arxiv": "2501.07572"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.07572"}], "name": "WebWalker: Benchmarking LLMs in Web Traversal", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-01-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebWalkerQA and an exploration-and-critique framework for gathering information by traversing multiple levels of a website.", "topics": ["evaluation", "web-navigation"], "updated_at": "2025-08-10", "url": "https://arxiv.org/abs/2501.07572"}
{"aliases": [], "authors": ["Boye Niu", "Yiliao Song", "Kai Lian", "Yifan Shen", "Yu Yao", "Kun Zhang", "Tongliang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57f5c4a37f662508b43896d42ed6d27bd5a7777fee98447b63e169cbe573109b", "submitted_at": "2025-01-14T04:35:37Z", "updated_at": "2025-02-23T06:20:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.07834v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2501.07834", "identifiers": {"arxiv": "2501.07834"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.07834"}], "name": "Flow: Modularized Agentic Workflow Automation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-23", "url": "https://arxiv.org/abs/2501.07834"}
{"aliases": [], "authors": ["Niklas Pfister", "Václav Volhejn", "Manuel Knott", "Santiago Arias", "Julia Bazińska", "Mykhailo Bichurin", "Alan Commike", "Janet Darling", "Peter Dienes", "Matthew Fiedler", "David Haber", "Matthias Kraft", "Marco Lancini", "Max Mathys", "Damián Pascual-Ortiz", "Jakub Podolak", "Adrià Romero-López", "Kyriacos Shiarlis", "Andreas Signer", "Zsolt Terek", "Athanasios Theocharis", "Daniel Timbrell", "Samuel Trautwein", "Samuel Watts", "Yun-Han Wu", "Mateo Rojas-Carulla"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1b1ccc6b13dd75414c880c0b83d79e43674a90be6c153e8165bfd10ff0825c77", "submitted_at": "2025-01-14T08:30:49Z", "updated_at": "2025-08-04T16:37:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.07927v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2501.07927", "identifiers": {"arxiv": "2501.07927"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.07927"}], "name": "Gandalf the Red: Adaptive Security for LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-04", "url": "https://arxiv.org/abs/2501.07927"}
{"aliases": [], "authors": ["Shaona Ghosh", "Prasoon Varshney", "Makesh Narsimhan Sreedhar", "Aishwarya Padmakumar", "Traian Rebedea", "Jibin Rajan Varghese", "Christopher Parisien"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40a3db0ea51d2b06ad914ae1444da5bb3c1fe511b9154c8c84866797849921e3", "submitted_at": "2025-01-15T18:37:08Z", "updated_at": "2025-01-15T18:37:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.09004v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2501.09004", "identifiers": {"arxiv": "2501.09004"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.09004"}], "name": "Aegis2.0: A Diverse AI Safety Dataset and Risks Taxonomy for Alignment of LLM Guardrails", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-15", "url": "https://arxiv.org/abs/2501.09004"}
{"aliases": [], "authors": ["Hongjin Su", "Ruoxi Sun", "Jinsung Yoon", "Pengcheng Yin", "Tao Yu", "Sercan Ö. Arık"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cd2a8cf1ddaced56f7a2a74ef2c1a6b8b0822b3fdbdde30796d08635b68484b0", "submitted_at": "2025-01-18T22:34:41Z", "updated_at": "2025-01-18T22:34:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.10893v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2501.10893", "identifiers": {"arxiv": "2501.10893"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.10893"}], "name": "Learn-by-interact: A Data-Centric Framework for Self-Adaptive Agents in Realistic Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-18", "url": "https://arxiv.org/abs/2501.10893"}
{"aliases": [], "authors": ["Kanika Goswami", "Puneet Mathur", "Ryan Rossi", "Franck Dernoncourt"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a1435686bf33522f817c5a542517465477ede5a3af89ff33625e737eb5a5c42f", "submitted_at": "2025-01-20T02:31:52Z", "updated_at": "2025-01-20T02:31:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.11233v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2501.11233", "identifiers": {"arxiv": "2501.11233"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.11233"}], "name": "PlotEdit: Natural Language-Driven Accessible Chart Editing in PDFs via Multimodal LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-20", "url": "https://arxiv.org/abs/2501.11233"}
{"aliases": [], "authors": ["Siyu Yuan", "Zehui Chen", "Zhiheng Xi", "Junjie Ye", "Zhengyin Du", "Jiecao Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "692a6984b464367c99beb9c5fbd4bbd8378e79f4a4b749503c8fbb72dd80b16f", "submitted_at": "2025-01-20T11:46:04Z", "updated_at": "2025-03-24T10:18:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.11425v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:06:43Z", "role": "primary", "sha256": "c35562ab132e13d69155842ec63a2f4189fe8703822e5620df2592a154d2d7ed", "supports": ["environments", "scope", "notes"], "url": "https://arxiv.org/html/2501.11425"}], "id": "arxiv:2501.11425", "identifiers": {"arxiv": "2501.11425"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.11425"}], "name": "Agent-R: Training Language Model Agents to Reflect via Iterative Self-Training", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The full paper was also inspected to confirm WebShop evaluation; the broader agent-learning method is classified as adjacent.", "published_at": "2025-01-20", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Builds self-critique training examples from searched recovery paths to improve agents' ability to detect and correct errors during interaction.", "topics": ["data-generation", "planning", "robustness"], "updated_at": "2025-03-24", "url": "https://arxiv.org/abs/2501.11425"}
{"aliases": [], "authors": ["Zhenhailong Wang", "Haiyang Xu", "Junyang Wang", "Xi Zhang", "Ming Yan", "Ji Zhang", "Fei Huang", "Heng Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5cf28f084cac7fc38e5c2dc4654115356d2d57f16009aad8925e071b8005be9c", "submitted_at": "2025-01-20T20:35:46Z", "updated_at": "2025-01-28T16:58:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.11733v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2501.11733", "identifiers": {"arxiv": "2501.11733"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.11733"}], "name": "Mobile-Agent-E: Self-Evolving Mobile Assistant for Complex Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-01-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses hierarchical mobile agents and persistent tips and reusable action shortcuts to improve complex tasks from accumulated experience.", "topics": ["mobile-navigation", "memory", "long-horizon"], "updated_at": "2025-01-28", "url": "https://arxiv.org/abs/2501.11733"}
{"aliases": [], "authors": ["Yujia Qin", "Yining Ye", "Junjie Fang", "Haoming Wang", "Shihao Liang", "Shizuo Tian", "Junda Zhang", "Jiahao Li", "Yunxin Li", "Shijue Huang", "Wanjun Zhong", "Kuanye Li", "Jiale Yang", "Yu Miao", "Woyu Lin", "Longxiang Liu", "Xu Jiang", "Qianli Ma", "Jingyu Li", "Xiaojun Xiao", "Kai Cai", "Chuang Li", "Yaowei Zheng", "Chaolin Jin", "Chen Li", "Xiao Zhou", "Minchao Wang", "Haoli Chen", "Zhaojian Li", "Haihua Yang", "Haifeng Liu", "Feng Lin", "Tao Peng", "Xin Liu", "Guang Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d19aa758e66411a31dd1fea6f07c83d1c828bd4aad86f1564a9d7f3670c3f79f", "submitted_at": "2025-01-21T17:48:10Z", "updated_at": "2025-01-21T17:48:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.12326v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2501.12326", "identifiers": {"arxiv": "2501.12326"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.12326"}], "name": "UI-TARS: Pioneering Automated GUI Interaction with Native Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-01-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies a screenshot-driven native GUI model with unified actions, deliberate reasoning, and iterative training on interaction traces.", "topics": ["grounding", "planning", "data-generation"], "updated_at": "2025-01-21", "url": "https://arxiv.org/abs/2501.12326"}
{"aliases": [], "authors": ["Yue Fan", "Handong Zhao", "Ruiyi Zhang", "Yu Shen", "Xin Eric Wang", "Gang Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0804055deb6c3513990ed8d596486400f1cb791338b7866d9e949186a34eb9c5", "submitted_at": "2025-01-23T18:16:21Z", "updated_at": "2025-01-27T18:58:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.13896v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2501.13896", "identifiers": {"arxiv": "2501.13896"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.13896"}], "name": "GUI-Bee: Align GUI Action Grounding to Novel Environments via Autonomous Exploration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-01-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Collects environment-specific interaction data through autonomous exploration to adapt grounding models to previously unseen GUIs.", "topics": ["grounding", "data-generation"], "updated_at": "2025-01-27", "url": "https://arxiv.org/abs/2501.13896"}
{"aliases": [], "authors": ["Lin Yueyu", "Li Zhiyuan", "Peter Yue", "Liu Xiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1b75157edc2f7afb538d4439f15a94b0f93e9beffbc328ca708fcfbaf0abbc18", "submitted_at": "2025-01-26T15:56:56Z", "updated_at": "2025-01-26T15:56:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.15570v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2501.15570", "identifiers": {"arxiv": "2501.15570"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.15570"}], "name": "ARWKV: Pretrain is not what we need, an RNN-Attention-Based Language Model Born from Transformer", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-26", "url": "https://arxiv.org/abs/2501.15570"}
{"aliases": [], "authors": ["Pascal J. Sager", "Benjamin Meyer", "Peng Yan", "Rebekka von Wartburg-Kottler", "Layan Etaiwi", "Aref Enayati", "Gabriel Nobel", "Ahmed Abdulkadir", "Benjamin F. Grewe", "Thilo Stadelmann"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b36fe818cb176c434b953fe8d2c97b67176b58cc6f8426b629231875de3c0896", "submitted_at": "2025-01-27T15:44:02Z", "updated_at": "2026-03-27T07:41:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.16150v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2501.16150", "identifiers": {"arxiv": "2501.16150"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.16150"}], "name": "A Comprehensive Survey of Agents for Computer Use: Foundations, Challenges, and Future Directions", "notes": "", "published_at": "2025-01-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Reviews 87 computer-use agents and 33 datasets through a taxonomy of operating domain, observation and action modalities, and agent design.", "topics": ["survey"], "updated_at": "2026-03-27", "url": "https://arxiv.org/abs/2501.16150"}
{"aliases": [], "authors": ["Domen Vake", "Bogdan Šinik", "Jernej Vičič", "Aleksandar Tošić"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4e5565347bb75bb7fd58c8f2e5c7e968f02386300d67225be335826d68c02671", "submitted_at": "2025-01-27T09:03:49Z", "updated_at": "2025-01-27T09:03:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.16403v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2501.16403", "identifiers": {"arxiv": "2501.16403"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.16403"}], "name": "Is Open Source the Future of AI? A Data-Driven Approach", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-01-27", "url": "https://arxiv.org/abs/2501.16403"}
{"aliases": [], "authors": ["Faria Huq", "Zora Zhiruo Wang", "Frank F. Xu", "Tianyue Ou", "Shuyan Zhou", "Jeffrey P. Bigham", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b14b88273749ac54ebbe3694cbe4196640da8f81ba2faf77b0e6bbd79a060f6", "submitted_at": "2025-01-28T00:56:53Z", "updated_at": "2026-02-27T05:35:06Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2501.16609v4"}, "publication": {"authors": ["Faria Huq", "Zora Zhiruo Wang", "Frank F. Xu", "Tianyue Ou", "Shuyan Zhou", "Jeffrey P. Bigham", "Graham Neubig"], "date_basis": "The ACL Anthology citation_publication_date is 2025/4, and its visible Month/Year fields state April 2025. Only month precision is supplied; no day is inferred from conference dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-demo.17", "event_dates": null, "evidence_urls": ["https://arxiv.org/abs/2501.16609v4", "https://aclanthology.org/2025.naacl-demo.17/"], "identity_basis": "The retained arXiv v4 abstract page explicitly links this DOI. The ACL publication has the same complete title and the same seven authors in the same order, and describes the same CowPilot human-agent navigation framework. This attaches the proceedings edition to the existing arXiv identity.", "issue": null, "notes": "The April 2025 proceedings edition is distinct from the retained arXiv v4, last revised on 2026-02-27. The publisher abstract has minor wording and typesetting differences from the retained arXiv abstract; neither its text nor its date replaces the original arXiv snapshot or annotation. Exact publication day, numeric volume, issue and series are not established.", "pages": "163-172", "pdf_url": "https://aclanthology.org/2025.naacl-demo.17.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "CowPilot: A Framework for Autonomous and Human-Agent Collaborative Web Navigation", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-demo.17/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:57Z", "role": "primary", "sha256": "8c39a7773a9f2b58950aa22d6924b00ed32ab5b2b42302590dcec88da255a71f", "supports": ["details.publication.doi", "details.publication.identity_basis", "details.publication.notes"], "url": "https://arxiv.org/abs/2501.16609v4"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:44:12Z", "role": "primary", "sha256": "410da7f4b7a79e56efc5f88014a078da0feb8d2ae3ef3d36a04dce5be29d4250", "supports": ["details.publication.type", "details.publication.title", "details.publication.authors", "details.publication.date_parts", "details.publication.date_basis", "details.publication.url", "details.publication.pdf_url", "details.publication.doi", "details.publication.venue", "details.publication.publisher", "details.publication.pages", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2025.naacl-demo.17/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:57Z", "role": "primary", "sha256": "8c39a7773a9f2b58950aa22d6924b00ed32ab5b2b42302590dcec88da255a71f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2501.16609v4"}], "id": "arxiv:2501.16609", "identifiers": {"arxiv": "2501.16609"}, "kind": "paper", "last_verified_at": "2026-10-17T11:19:57Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.16609"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-demo.17/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-demo.17.pdf"}], "name": "CowPilot: A Framework for Autonomous and Human-Agent Collaborative Web Navigation", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-01-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CowPilot, a web-navigation framework where agents suggest actions and users can pause, reject, override, or resume them. It evaluates autonomous and collaborative use through task success and human effort.", "topics": ["web-navigation", "human-in-the-loop", "evaluation"], "updated_at": "2026-02-27", "url": "https://arxiv.org/abs/2501.16609"}
{"aliases": [], "authors": ["Zihui Zhao", "Yingxin Li", "Yang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "796fa42be43e80e9680f0145f6a3327ba6ab12eaebbc9a4a1d8ec5e994ef8826", "submitted_at": "2025-01-29T02:52:32Z", "updated_at": "2025-09-30T10:14:30Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.17391v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2501.17391", "identifiers": {"arxiv": "2501.17391"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2501.17391"}], "name": "LFTR: Learning-Free Token Reduction for Multimodal Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-01-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-30", "url": "https://arxiv.org/abs/2501.17391"}
{"aliases": [], "authors": ["Zhixi Cai", "Fucai Ke", "Simindokht Jahangard", "Maria Garcia de la Banda", "Reza Haffari", "Peter J. Stuckey", "Hamid Rezatofighi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "24e828daaede60a8714bb4c1ee20810389efa3faf372a9f6bc106443a5fba066", "submitted_at": "2025-02-01T09:19:08Z", "updated_at": "2025-08-14T12:58:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.00372v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.00372", "identifiers": {"arxiv": "2502.00372"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.00372"}], "name": "NAVER: A Neuro-Symbolic Compositional Automaton for Visual Grounding with Explicit Logic Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-14", "url": "https://arxiv.org/abs/2502.00372"}
{"aliases": [], "authors": ["Zhichao Sun", "Yepeng Liu", "Zhiling Su", "Huachao Zhu", "Yuliang Gu", "Yuda Zou", "Zelong Liu", "Gui-Song Xia", "Bo Du", "Yongchao Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "832b3fea3b7239a80ce7a2b861edde162fd196c39c307c5ac09a3b0ee21f0b6e", "submitted_at": "2025-02-01T10:44:11Z", "updated_at": "2025-11-24T02:48:47Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.00392v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.00392", "identifiers": {"arxiv": "2502.00392"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.00392"}], "name": "RefDrone: A Challenging Benchmark for Referring Expression Comprehension in Drone Scenes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-24", "url": "https://arxiv.org/abs/2502.00392"}
{"aliases": [], "authors": ["Harshith Padigela", "Chintan Shah", "Dinkar Juyal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0ce1cbe36e927c7ea4b300f4f59cb0a454ab65aced9449f15a3ea6a11b87ac3", "submitted_at": "2025-02-03T00:04:49Z", "updated_at": "2025-02-19T05:09:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.00964v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.00964", "identifiers": {"arxiv": "2502.00964"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.00964"}], "name": "ML-Dev-Bench: Comparative Analysis of AI Agents on ML development workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-19", "url": "https://arxiv.org/abs/2502.00964"}
{"aliases": [], "authors": ["Kanika Goswami", "Puneet Mathur", "Ryan Rossi", "Franck Dernoncourt"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e94a478502b990bcd58492dea39d1d6b7ecfddeb06c21db75c8d55e5814e4b62", "submitted_at": "2025-02-03T02:00:29Z", "updated_at": "2025-02-03T02:00:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.00988v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.00988", "identifiers": {"arxiv": "2502.00988"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.00988"}], "name": "PlotGen: Multi-Agent LLM-based Scientific Data Visualization via Multimodal Feedback", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-03", "url": "https://arxiv.org/abs/2502.00988"}
{"aliases": [], "authors": ["Minttu Alakuijala", "Ya Gao", "Georgy Ananov", "Samuel Kaski", "Pekka Marttinen", "Alexander Ilin", "Harri Valpola"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "907cc7ed7ac7651384af88703981d3136db68d66da38c4b47e90037ac8266aa7", "submitted_at": "2025-02-03T17:45:46Z", "updated_at": "2026-09-20T18:47:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.01562v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.01562", "identifiers": {"arxiv": "2502.01562"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.01562"}], "name": "Memento No More: Coaching AI Agents to Master Multiple Tasks via Hints Internalization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-20", "url": "https://arxiv.org/abs/2502.01562"}
{"aliases": [], "authors": ["Kevin Chen", "Marco Cusumano-Towner", "Brody Huval", "Aleksei Petrenko", "Jackson Hamburger", "Vladlen Koltun", "Philipp Krähenbühl"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e04fe57f9bc51d03f3fff4bf177ff512316ee9df20aaea034662f3ea3104cd97", "submitted_at": "2025-02-03T18:35:42Z", "updated_at": "2025-03-08T05:23:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.01600v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2502.01600", "identifiers": {"arxiv": "2502.01600"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.01600"}], "name": "Reinforcement Learning for Long-Horizon Interactive LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-08", "url": "https://arxiv.org/abs/2502.01600"}
{"aliases": [], "authors": ["Eric Zhao", "Pranjal Awasthi", "Sreenivas Gollapudi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b26677a629a9edb01f14dcccd469d9ff309a2a3d1f7be13833f6115e970031f", "submitted_at": "2025-02-03T21:31:07Z", "updated_at": "2025-02-20T18:52:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.01839v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.01839", "identifiers": {"arxiv": "2502.01839"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.01839"}], "name": "Sample, Scrutinize and Scale: Effective Inference-Time Search by Scaling Verification", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-20", "url": "https://arxiv.org/abs/2502.01839"}
{"aliases": [], "authors": ["Hongxin Li", "Jingfan Chen", "Jingran Su", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b76ab1c6c44db9f86b2bfc11097c54e5c3115c0c498a58d896bef043d2113c8", "submitted_at": "2025-02-04T03:39:59Z", "updated_at": "2025-06-07T05:23:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.01977v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2502.01977", "identifiers": {"arxiv": "2502.01977"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.01977"}], "name": "AutoGUI: Scaling GUI Grounding with Automatic Functionality Annotations from LLMs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Infers element functionality from UI changes before and after interaction to produce verified annotations for grounding-model training.", "topics": ["grounding", "data-generation"], "updated_at": "2025-06-07", "url": "https://arxiv.org/abs/2502.01977"}
{"aliases": [], "authors": ["Han Zhou", "Xingchen Wan", "Ruoxi Sun", "Hamid Palangi", "Shariq Iqbal", "Ivan Vulić", "Anna Korhonen", "Sercan Ö. Arık"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8344d1f0c0cd6f82704d664e0ce563b72f38dfc37494b1c47aec962e679ed69a", "submitted_at": "2025-02-04T17:56:44Z", "updated_at": "2026-01-31T22:11:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.02533v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2502.02533", "identifiers": {"arxiv": "2502.02533"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.02533"}], "name": "Multi-Agent Design: Optimizing Agents with Better Prompts and Topologies", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-31", "url": "https://arxiv.org/abs/2502.02533"}
{"aliases": [], "authors": ["Qinzhuo Wu", "Wei Liu", "Jian Luan", "Bin Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3e0b19f259bcdb57b0dea92578d0ee1c7c229ea11e622c56e808f17ca2568cc", "submitted_at": "2025-02-05T07:35:23Z", "updated_at": "2025-02-05T07:35:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.02955v1"}, "publication": {"authors": ["Qinzhuo Wu", "Wei Liu", "Jian Luan", "Bin Wang"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-long.244", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-long.244/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "4760-4775", "pdf_url": "https://aclanthology.org/2025.naacl-long.244.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "ReachAgent: Enhancing Mobile Agent via Page Reaching and Operation", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-long.244/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:58Z", "role": "primary", "sha256": "bb33865bafca385d3d1ce446d5cd85b6c1bf41d567509ac5ea539d4cc9c64799", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-long.244/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2502.02955", "identifiers": {"arxiv": "2502.02955"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.02955"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-long.244/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-long.244.pdf"}], "name": "ReachAgent: Enhancing Mobile Agent via Page Reaching and Operation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Splits mobile tasks into reaching the required page and operating on it, using subtask data and preferred GUI trajectories for training.", "topics": ["mobile-navigation", "planning", "data-generation"], "updated_at": "2025-02-05", "url": "https://arxiv.org/abs/2502.02955"}
{"aliases": [], "authors": ["Wenhao Wang", "Mengying Yuan", "Zijie Yu", "Guangyi Liu", "Rui Ye", "Tian Jin", "Siheng Chen", "Yanfeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecab608111104ac044bb7d28996604e9b87bdad4543c250c57bf432f883b96d5", "submitted_at": "2025-02-05T08:26:17Z", "updated_at": "2025-05-20T07:03:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.02982v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2502.02982", "identifiers": {"arxiv": "2502.02982"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.02982"}], "name": "MobileA3gent: Training Mobile GUI Agents Using Decentralized Self-Sourced Data from Diverse Users", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies automatic annotation and federated training of mobile GUI agents from decentralized user interaction data with heterogeneous distributions.", "topics": ["mobile-navigation", "data-generation"], "updated_at": "2025-05-20", "url": "https://arxiv.org/abs/2502.02982"}
{"aliases": [], "authors": ["Muhan Lin", "Shuyang Shi", "Yue Guo", "Vaishnav Tadiparthi", "Behdad Chalaki", "Ehsan Moradi Pari", "Simon Stepputtis", "Woojun Kim", "Joseph Campbell", "Katia Sycara"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db318b3f2bcbbb87f6b1a932fdb0182978544317413098034e1dc67420780916", "submitted_at": "2025-02-06T02:26:47Z", "updated_at": "2025-03-01T01:43:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.03723v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.03723", "identifiers": {"arxiv": "2502.03723"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.03723"}], "name": "Speaking the Language of Teamwork: LLM-Guided Credit Assignment in Multi-Agent Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-01", "url": "https://arxiv.org/abs/2502.03723"}
{"aliases": [], "authors": ["Kaijie Zhu", "Xianjun Yang", "Jindong Wang", "Wenbo Guo", "William Yang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3184681ee3691a46232c9922c6ef9882401425871237c498f06e76c71495c426", "submitted_at": "2025-02-07T18:57:49Z", "updated_at": "2025-06-10T18:13:09Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2502.05174v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.05174", "identifiers": {"arxiv": "2502.05174"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.05174"}], "name": "MELON: Provable Defense Against Indirect Prompt Injection Attacks in AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-10", "url": "https://arxiv.org/abs/2502.05174"}
{"aliases": [], "authors": ["Gonzalo Gonzalez-Pumariega", "Leong Su Yean", "Neha Sunkara", "Sanjiban Choudhury"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4323d5e7493c0a8f0f56b80f1506a7c839665ce1bc9e45f18243b99eb94e6157", "submitted_at": "2025-02-06T05:50:37Z", "updated_at": "2025-02-06T05:50:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.05227v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.05227", "identifiers": {"arxiv": "2502.05227"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.05227"}], "name": "Robotouille: An Asynchronous Planning Benchmark for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-06", "url": "https://arxiv.org/abs/2502.05227"}
{"aliases": [], "authors": ["Georgios Papoudakis", "Thomas Coste", "Zhihao Wu", "Jianye Hao", "Jun Wang", "Kun Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "44ec9907c8d536488e59dea307aa05bd8adcadf96cb4f3ba25a355ffe4cf8ff7", "submitted_at": "2025-02-10T12:32:21Z", "updated_at": "2025-02-10T12:32:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.06395v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2502.06395", "identifiers": {"arxiv": "2502.06395"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.06395"}], "name": "AppVLM: A Lightweight Vision Language Model for Online App Control", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "AppVLM first learns app actions from AndroidControl and then refines its policy through data collected in AndroidWorld. The study compares offline action prediction and online task completion for a lightweight vision-language smartphone agent.", "topics": ["mobile-navigation", "imitation-learning", "data-generation", "efficiency"], "updated_at": "2025-02-10", "url": "https://arxiv.org/abs/2502.06395"}
{"aliases": [], "authors": ["Ruichen Zhang", "Mufan Qiu", "Zhen Tan", "Mohan Zhang", "Vincent Lu", "Jie Peng", "Kaidi Xu", "Leandro Z. Agudelo", "Peter Qian", "Tianlong Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0535de4778b6a31adb64f833733ebf94acf54ba872d98afd40b3e1012fb737fd", "submitted_at": "2025-02-11T20:41:49Z", "updated_at": "2025-03-06T19:40:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.07942v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2502.07942", "identifiers": {"arxiv": "2502.07942"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.07942"}], "name": "Symbiotic Cooperation for Web Agents: Harnessing Complementary Strengths of Large and Small LLMs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Couples large- and small-model web-agent exploration with trajectory synthesis and distillation to improve both model types iteratively.", "topics": ["data-generation", "web-navigation", "efficiency"], "updated_at": "2025-03-06", "url": "https://arxiv.org/abs/2502.07942"}
{"aliases": [], "authors": ["Qingyuan Wu", "Jianheng Liu", "Jianye Hao", "Jun Wang", "Kun Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "186b7fd8e79db290ea481645d8adab3989a12f787a13ecb11270f4756f5af42a", "submitted_at": "2025-02-11T20:57:46Z", "updated_at": "2025-05-20T19:54:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.07949v2"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2502.07949", "identifiers": {"arxiv": "2502.07949"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.07949"}], "name": "Advancing Autonomous VLM Agents via Variational Subgoal-Conditioned Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses variational subgoal-conditioned reinforcement learning to improve learning efficiency on sparse-reward mobile and web-control tasks.", "topics": ["reinforcement-learning", "planning"], "updated_at": "2025-05-20", "url": "https://arxiv.org/abs/2502.07949"}
{"aliases": [], "authors": ["Henry Hengyuan Zhao", "Kaiming Yang", "Wendi Yu", "Difei Gao", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68e6ecfcdff0163debe96f6440fb33529356bf9fd2c8743b5953ae52ae71c9ef", "submitted_at": "2025-02-12T01:06:10Z", "updated_at": "2026-05-23T20:19:43Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2502.08047v5"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2502.08047", "identifiers": {"arxiv": "2502.08047"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.08047"}], "name": "WorldGUI: An Interactive Benchmark for Desktop GUI Automation from Any Starting Point", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces desktop and web tasks with varied initial states to test planning recovery and adaptability beyond canonical starting configurations.", "topics": ["evaluation", "robustness", "planning"], "updated_at": "2026-05-23", "url": "https://arxiv.org/abs/2502.08047"}
{"aliases": [], "authors": ["Kunal Singh", "Shreyas Singh", "Mukund Khanna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a60d581e4318db66495a0333b864757ec896576c720e40da4a881f8740b9d9cf", "submitted_at": "2025-02-12T09:12:30Z", "updated_at": "2025-02-14T06:23:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.08226v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2502.08226", "identifiers": {"arxiv": "2502.08226"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.08226"}], "name": "TRISHUL: Towards Region Identification and Screen Hierarchy Understanding for Large VLM based GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines hierarchical screen parsing with spatially enriched element descriptions to support both locating GUI targets and describing indicated elements.", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-02-14", "url": "https://arxiv.org/abs/2502.08226"}
{"aliases": [], "authors": ["Aitor Gonzalez-Agirre", "Marc Pàmies", "Joan Llop", "Irene Baucells", "Severino Da Dalt", "Daniel Tamayo", "José Javier Saiz", "Ferran Espuña", "Jaume Prats", "Javier Aula-Blasco", "Mario Mina", "Iñigo Pikabea", "Adrián Rubio", "Alexander Shvets", "Anna Sallés", "Iñaki Lacunza", "Jorge Palomar", "Júlia Falcão", "Lucía Tormo", "Luis Vasquez-Reina", "Montserrat Marimon", "Oriol Pareras", "Valle Ruiz-Fernández", "Marta Villegas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f5c89b83cccd5d82fbbcceeafe77c3f0f4574254553fac55b5851155b8cc611", "submitted_at": "2025-02-12T15:26:08Z", "updated_at": "2025-02-13T17:33:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.08489v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.08489", "identifiers": {"arxiv": "2502.08489"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.08489"}], "name": "Salamandra Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-13", "url": "https://arxiv.org/abs/2502.08489"}
{"aliases": [], "authors": ["Hui Wei", "Zihao Zhang", "Shenghua He", "Tian Xia", "Shijia Pan", "Fei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27656a74c7807df12e1f8e5f831b6df7b4640d833082bc58af01b810d2f7f1fc", "submitted_at": "2025-02-16T17:54:57Z", "updated_at": "2025-06-23T05:32:12Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.11221v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.11221", "identifiers": {"arxiv": "2502.11221"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.11221"}], "name": "PlanGenLLMs: A Modern Survey of LLM Planning Capabilities", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-23", "url": "https://arxiv.org/abs/2502.11221"}
{"aliases": [], "authors": ["Vardaan Pahuja", "Yadong Lu", "Corby Rosset", "Boyu Gou", "Arindam Mitra", "Spencer Whitehead", "Yu Su", "Ahmed Awadallah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d9f459a02b7a464d9e1e49579943d1b93d46308c1b00e795c903d2fa1502a2b2", "submitted_at": "2025-02-17T02:13:48Z", "updated_at": "2025-05-30T02:30:57Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2502.11357v4"}, "publication": {"authors": ["Vardaan Pahuja", "Yadong Lu", "Corby Rosset", "Boyu Gou", "Arindam Mitra", "Spencer Whitehead", "Yu Su", "Ahmed Awadallah"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.326", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.326/", "https://aclanthology.org/2025.findings-acl.326.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the publisher PDF, page 1. Landing-page citation metadata gives Ahmed Hassan for the final author, while the supplied BibTeX expands the name to Ahmed Hassan Awadallah; the PDF uses Ahmed Awadallah.", "pages": "6300-6323", "pdf_url": "https://aclanthology.org/2025.findings-acl.326.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Explorer: Scaling Exploration-driven Web Trajectory Synthesis for Multimodal Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.326/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:58Z", "role": "primary", "sha256": "8ec40afe322b7c72c9f6d6db617fa2d4fae559cf6cd4e2fcc0ff5a9cc9485e05", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.326/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:02.000475+00:00", "role": "primary", "sha256": "c1b232e161839514713789c408ae64f66684b49770033c7f3b04bf1d1a232157", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2025.findings-acl.326.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2502.11357", "identifiers": {"arxiv": "2502.11357"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.11357"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.326/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.326.pdf"}], "name": "Explorer: Scaling Exploration-driven Web Trajectory Synthesis for Multimodal Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Synthesizes multimodal web trajectories through exploration and refinement, then trains a browser agent to study the effect of scaling interaction data.", "topics": ["data-generation", "web-navigation"], "updated_at": "2025-05-30", "url": "https://arxiv.org/abs/2502.11357"}
{"aliases": [], "authors": ["Wujiang Xu", "Zujie Liang", "Kai Mei", "Hang Gao", "Juntao Tan", "Yongfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d301f1761cacf9d4be63256126070f29bcfd02d4769499764bee812b424c5d6", "submitted_at": "2025-02-17T18:36:14Z", "updated_at": "2025-10-08T01:46:37Z", "version": 11, "versioned_url": "https://arxiv.org/abs/2502.12110v11"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2502.12110", "identifiers": {"arxiv": "2502.12110"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.12110"}], "name": "A-MEM: Agentic Memory for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-08", "url": "https://arxiv.org/abs/2502.12110"}
{"aliases": [], "authors": ["Zhenfang Chen", "Delin Chen", "Rui Sun", "Wenjun Liu", "Chuang Gan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5cde1d02d18500109d1190869f27eadc5f14d56fd09507d7a415e462c66aca6e", "submitted_at": "2025-02-17T18:49:25Z", "updated_at": "2025-02-17T18:49:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.12130v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.12130", "identifiers": {"arxiv": "2502.12130"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.12130"}], "name": "Scaling Autonomous Agents via Automatic Reward Modeling And Planning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-17", "url": "https://arxiv.org/abs/2502.12130"}
{"aliases": [], "authors": ["Yurun Chen", "Xavier Hu", "Keting Yin", "Juncheng Li", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5af8d3a8b4db150bab0630f6040a0f3bced612f68b11d31985053f1f25c0f6f6", "submitted_at": "2025-02-18T17:01:28Z", "updated_at": "2025-08-06T07:04:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.13053v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2502.13053", "identifiers": {"arxiv": "2502.13053"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.13053"}], "name": "Evaluating the Robustness of Multimodal Agents Against Active Environmental Injection Attacks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Examines active environmental injection against mobile agents, testing whether adversarial interface content redirects decisions during task execution.", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2025-08-06", "url": "https://arxiv.org/abs/2502.13053"}
{"aliases": [], "authors": ["Jianwei Yang", "Reuben Tan", "Qianhui Wu", "Ruijie Zheng", "Baolin Peng", "Yongyuan Liang", "Yu Gu", "Mu Cai", "Seonghyeon Ye", "Joel Jang", "Yuquan Deng", "Lars Liden", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d126515d15bc1d6749a2f2fc0419b6812906524562b5877b690be751ff4f61a", "submitted_at": "2025-02-18T18:55:21Z", "updated_at": "2025-02-18T18:55:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.13130v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T08:01:28Z", "role": "primary", "sha256": "822cec31fa0db7ffa390fda806d88c2f9c0fe4d328742e7008b662916e92373a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2005.07648%2C2206.04615%2C2311.12022%2C2311.16922%2C2502.13130%2C2506.21506%2C2511.11332&max_results=7"}], "id": "arxiv:2502.13130", "identifiers": {"arxiv": "2502.13130"}, "kind": "paper", "last_verified_at": "2026-10-18T08:01:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.13130"}], "name": "Magma: A Foundation Model for Multimodal AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-18", "url": "https://arxiv.org/abs/2502.13130"}
{"aliases": [], "authors": ["Shuai Bai", "Keqin Chen", "Xuejing Liu", "Jialin Wang", "Wenbin Ge", "Sibo Song", "Kai Dang", "Peng Wang", "Shijie Wang", "Jun Tang", "Humen Zhong", "Yuanzhi Zhu", "Mingkun Yang", "Zhaohai Li", "Jianqiang Wan", "Pengfei Wang", "Wei Ding", "Zheren Fu", "Yiheng Xu", "Jiabo Ye", "Xi Zhang", "Tianbao Xie", "Zesen Cheng", "Hang Zhang", "Zhibo Yang", "Haiyang Xu", "Junyang Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da29b5cb2aa57ccdde2a365faa41657e05d0017ea0f932664f3a06a6dcdfb8c6", "submitted_at": "2025-02-19T18:00:14Z", "updated_at": "2025-02-19T18:00:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.13923v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2502.13923", "identifiers": {"arxiv": "2502.13923"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.13923"}], "name": "Qwen2.5-VL Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-19", "url": "https://arxiv.org/abs/2502.13923"}
{"aliases": [], "authors": ["Haowei Liu", "Xi Zhang", "Haiyang Xu", "Yuyang Wanyan", "Junyang Wang", "Ming Yan", "Ji Zhang", "Chunfeng Yuan", "Changsheng Xu", "Weiming Hu", "Fei Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d75b417f4248397b73934f8b8e45199006e4db6e2ee69b62e0846621411b2dd7", "submitted_at": "2025-02-20T05:41:55Z", "updated_at": "2025-02-21T02:54:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.14282v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2502.14282", "identifiers": {"arxiv": "2502.14282"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.14282"}], "name": "PC-Agent: A Hierarchical Multi-Agent Collaboration Framework for Complex Task Automation on PC", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses active perception and hierarchical agents for instruction decomposition, progress tracking, action selection, and reflection in PC workflows.", "topics": ["planning", "cross-app"], "updated_at": "2025-02-21", "url": "https://arxiv.org/abs/2502.14282"}
{"aliases": [], "authors": ["Zhitao He", "Zijun Liu", "Peng Li", "Yi R. Fung", "Ming Yan", "Ji Zhang", "Fei Huang", "Yang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9dc80d2365e9134283f2094edb5f4d1025d32617b45f33320edb4d825264b318", "submitted_at": "2025-02-20T12:26:15Z", "updated_at": "2025-08-20T09:42:24Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.14496v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.14496", "identifiers": {"arxiv": "2502.14496"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.14496"}], "name": "Advancing Language Multi-Agent Learning with Credit Re-Assignment for Interactive Environment Generalization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-20", "url": "https://arxiv.org/abs/2502.14496"}
{"aliases": [], "authors": ["Hao Bai", "Yifei Zhou", "Li Erran Li", "Sergey Levine", "Aviral Kumar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "99492dab54d9816104d47fca525277b033cadb7a53727dae06ed1a3221fcb416", "submitted_at": "2025-02-13T18:55:14Z", "updated_at": "2025-02-13T18:55:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.15760v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2502.15760", "identifiers": {"arxiv": "2502.15760"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.15760"}], "name": "Digi-Q: Learning Q-Value Functions for Training Device-Control Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Learns action-value functions from offline device-control experience and uses them to rank candidate actions for policy improvement.", "topics": ["reinforcement-learning", "efficiency"], "updated_at": "2025-02-13", "url": "https://arxiv.org/abs/2502.15760"}
{"aliases": [], "authors": ["Sicheng Xie", "Haidong Cao", "Zejia Weng", "Zhen Xing", "Haoran Chen", "Shiwei Shen", "Jiaqi Leng", "Zuxuan Wu", "Yu-Gang Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6392da545f75c3799addb56d83ca3af495f837ae73d16cbcc66b3567db6b3414", "submitted_at": "2025-02-23T14:29:28Z", "updated_at": "2025-11-15T06:58:42Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2502.16587v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.16587", "identifiers": {"arxiv": "2502.16587"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.16587"}], "name": "Human2Robot: Learning Robot Actions from Paired Human-Robot Videos", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-15", "url": "https://arxiv.org/abs/2502.16587"}
{"aliases": [], "authors": ["Liangtao Shi", "Ting Liu", "Xiantao Hu", "Yue Hu", "Quanjun Yin", "Richang Hong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "161ec91078c2cd27cc2dca4ab9c0b506cea930743ec332f860fface907dc47e2", "submitted_at": "2025-02-24T02:41:34Z", "updated_at": "2025-02-28T12:54:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.16786v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.16786", "identifiers": {"arxiv": "2502.16786"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.16786"}], "name": "SwimVG: Step-wise Multimodal Fusion and Adaption for Visual Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-28", "url": "https://arxiv.org/abs/2502.16786"}
{"aliases": [], "authors": ["Yuxuan Liu", "Hongda Sun", "Wei Liu", "Jian Luan", "Bo Du", "Rui Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3522393279755638a2ca417e42a664f17fdad6991b9ce1103f2ef792090d4966", "submitted_at": "2025-02-24T03:12:45Z", "updated_at": "2025-02-24T03:12:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.16796v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.16796", "identifiers": {"arxiv": "2502.16796"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.16796"}], "name": "MobileSteward: Integrating Multiple App-Oriented Agents with Self-Evolution to Automate Cross-App Instructions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-24", "url": "https://arxiv.org/abs/2502.16796"}
{"aliases": [], "authors": ["Junyang Wang", "Haiyang Xu", "Xi Zhang", "Ming Yan", "Ji Zhang", "Fei Huang", "Jitao Sang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d723cb764deb78b194ff0199c7f1b921a12b6a41655c07b06ae88c3d0af58358", "submitted_at": "2025-02-24T12:51:23Z", "updated_at": "2025-06-03T08:37:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.17110v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2502.17110", "identifiers": {"arxiv": "2502.17110"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.17110"}], "name": "Mobile-Agent-V: A Video-Guided Approach for Effortless and Efficient Operational Knowledge Injection in Mobile Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The authors identify arXiv:2505.13887 as an accidental duplicate and direct future updates to this record.", "published_at": "2025-02-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Extracts operational guidance from video for mobile task execution and introduces Mobile-Knowledge to evaluate the contribution of external knowledge.", "topics": ["mobile-navigation", "video", "memory"], "updated_at": "2025-06-03", "url": "https://arxiv.org/abs/2502.17110"}
{"aliases": [], "authors": ["Lars Krupp", "Daniel Geißler", "Paul Lukowicz", "Jakob Karolus"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3903e97e6db6bd470699f0d7b2785e53e15291234f1e5f717bae8d2ccd851ca4", "submitted_at": "2025-02-25T06:58:40Z", "updated_at": "2025-02-25T06:58:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.17903v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2502.17903", "identifiers": {"arxiv": "2502.17903"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.17903"}], "name": "Towards Sustainable Web Agents: A Plea for Transparency and Dedicated Metrics for Energy Consumption", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Examines energy and carbon-accounting limitations in web-agent evaluation and argues for transparent, dedicated sustainability measurements.", "topics": ["evaluation", "efficiency"], "updated_at": "2025-02-25", "url": "https://arxiv.org/abs/2502.17903"}
{"aliases": [], "authors": ["George Thomas", "Alex J. Chan", "Jikun Kang", "Wenqi Wu", "Filippos Christianos", "Fraser Greenlee", "Andy Toulis", "Marvin Purtorab"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c21fbd381caec6348dfdc183a57331e316d8747715425843451b563b3ad17999", "submitted_at": "2025-02-25T16:45:08Z", "updated_at": "2025-02-25T16:45:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.18356v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.18356", "identifiers": {"arxiv": "2502.18356"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.18356"}], "name": "WebGames: Challenging General-Purpose Web-Browsing AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-25", "url": "https://arxiv.org/abs/2502.18356"}
{"aliases": [], "authors": ["Jianjian Li", "Junquan Fan", "Feng Tang", "Gang Huang", "Shitao Zhu", "Songlin Liu", "Nian Xie", "Wulong Liu", "Yong Liao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e537c9c85fdd0a1f3ca21dd44421151c00ac91c41cca3ce4543ccdf62ae44e6a", "submitted_at": "2025-02-22T16:05:33Z", "updated_at": "2025-02-22T16:05:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.18512v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.18512", "identifiers": {"arxiv": "2502.18512"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.18512"}], "name": "FCoT-VL:Advancing Text-oriented Large Vision-Language Models with Efficient Visual Token Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-22", "url": "https://arxiv.org/abs/2502.18512"}
{"aliases": [], "authors": ["Pranjal Aggarwal", "Sean Welleck"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e4091fc740249b6406dd9e96f2692d7fac9323759ec6fad6ba5a05f3e445844", "submitted_at": "2025-02-24T18:41:33Z", "updated_at": "2025-10-03T02:09:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.18525v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2502.18525", "identifiers": {"arxiv": "2502.18525"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.18525"}], "name": "Programming with Pixels: Can Computer-Use Agents do Software Engineering?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a visual IDE environment and benchmark for software-engineering tasks, comparing screenshot-based control with access to editing and shell tools.", "topics": ["evaluation", "tool-integration", "automation"], "updated_at": "2025-10-03", "url": "https://arxiv.org/abs/2502.18525"}
{"aliases": [], "authors": ["Jiani Zheng", "Lu Wang", "Fangkai Yang", "Chaoyun Zhang", "Lingrui Mei", "Wenjie Yin", "Qingwei Lin", "Dongmei Zhang", "Saravan Rajmohan", "Qi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d69b006a298a591aacfe7f29df63f508113c0ec824ede82a0bb5bbb9cb66266", "submitted_at": "2025-02-26T07:52:02Z", "updated_at": "2025-02-26T07:52:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.18906v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2502.18906", "identifiers": {"arxiv": "2502.18906"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.18906"}], "name": "VEM: Environment-Free Exploration for Training GUI Agent with Value Environment Model", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains GUI agents with a frozen value model learned from offline interactions, guiding action exploration without predicting complete next observations.", "topics": ["reinforcement-learning", "world-models"], "updated_at": "2025-02-26", "url": "https://arxiv.org/abs/2502.18906"}
{"aliases": [], "authors": ["Hao Peng", "Yunjia Qi", "Xiaozhi Wang", "Zijun Yao", "Bin Xu", "Lei Hou", "Juanzi Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a454d3baa94389440f866b71013ab7a763a092879962ceb3c5978e86a7740b0", "submitted_at": "2025-02-26T17:19:12Z", "updated_at": "2025-02-26T17:19:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.19328v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2502.19328", "identifiers": {"arxiv": "2502.19328"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.19328"}], "name": "Agentic Reward Modeling: Integrating Human Preferences with Verifiable Correctness Signals for Reliable Reward Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-26", "url": "https://arxiv.org/abs/2502.19328"}
{"aliases": [], "authors": ["Jeffrey Yang Fan Chiang", "Seungjae Lee", "Jia-Bin Huang", "Furong Huang", "Yizheng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27eacc2668411e2b7f79423940aa12293fa57aa81390c64587d3074c141dcb4f", "submitted_at": "2025-02-27T18:56:26Z", "updated_at": "2025-09-21T22:29:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.20383v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:04Z", "role": "primary", "sha256": "0118540c95d316ed30aabf783c072d413677aaf89539316e08e1d81ddab2c5e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2502.20383v3"}], "id": "arxiv:2502.20383", "identifiers": {"arxiv": "2502.20383"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.20383"}], "name": "Why Are Web AI Agents More Vulnerable Than Standalone LLMs? A Security Analysis", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-02-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes web-agent vulnerabilities at the component level, examining goal placement in prompts, multistep action generation, and observations as sources of increased susceptibility.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2025-09-21", "url": "https://arxiv.org/abs/2502.20383"}
{"aliases": [], "authors": ["Sirui Xu", "Hung Yu Ling", "Yu-Xiong Wang", "Liang-Yan Gui"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15589f068685934662d53fb17eb4dee09efb46e20137a8288d5d431380eee774", "submitted_at": "2025-02-27T18:59:12Z", "updated_at": "2026-02-01T18:39:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.20390v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.20390", "identifiers": {"arxiv": "2502.20390"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.20390"}], "name": "InterMimic: Towards Universal Whole-Body Control for Physics-Based Human-Object Interactions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-01", "url": "https://arxiv.org/abs/2502.20390"}
{"aliases": [], "authors": ["Javin Liu", "Aryan Vats", "Zihao He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "92246b1f3e3f6f836b50013fc6be343e8370a7b49da60ceea3236d3adf88c38f", "submitted_at": "2025-02-27T22:48:35Z", "updated_at": "2025-02-27T22:48:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.20582v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2502.20582", "identifiers": {"arxiv": "2502.20582"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2502.20582"}], "name": "CS-PaperSum: A Large-Scale Dataset of AI-Generated Summaries for Scientific Papers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-02-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-02-27", "url": "https://arxiv.org/abs/2502.20582"}
{"aliases": [], "authors": ["Zongru Wu", "Pengzhou Cheng", "Zheng Wu", "Tianjie Ju", "Zhuosheng Zhang", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53781edbb107437b258d0c0e298f80ec14f8f32b51bc0dd11c0417473eb2c606", "submitted_at": "2025-03-01T08:29:59Z", "updated_at": "2025-03-04T12:04:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.00401v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.00401", "identifiers": {"arxiv": "2503.00401"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.00401"}], "name": "Smoothing Grounding and Reasoning for MLLM-Powered GUI Agents with Query-Oriented Pivot Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses inference of possible user queries from screenshots and target coordinates as a training bridge between GUI grounding and action reasoning.", "topics": ["grounding", "planning"], "updated_at": "2025-03-04", "url": "https://arxiv.org/abs/2503.00401"}
{"aliases": [], "authors": ["Sami Marreed", "Alon Oved", "Avi Yaeli", "Segev Shlomov", "Ido Levy", "Offer Akrabi", "Aviad Sela", "Asaf Adi", "Nir Mashkif"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2248d423c954ee6e406e3f46f6da43fabd6e832ffacc05cfa32083f2ff0c2700", "submitted_at": "2025-02-24T09:31:56Z", "updated_at": "2025-07-09T13:52:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.01861v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.01861", "identifiers": {"arxiv": "2503.01861"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.01861"}], "name": "Towards Enterprise-Ready Computer Using Generalist Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes the CUGA enterprise-agent development process, emphasizing iterative evaluation, failure analysis, and refinement on software-task benchmarks.", "topics": ["automation", "evaluation", "workflow"], "updated_at": "2025-07-09", "url": "https://arxiv.org/abs/2503.01861"}
{"aliases": [], "authors": ["Zhixun Chen", "Ming Li", "Yuxuan Huang", "Yali Du", "Meng Fang", "Tianyi Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a58e33cd2b60a8c734968dd92f031b9aa4811932b1a666d947b20ec4c6035a7", "submitted_at": "2025-03-04T02:14:55Z", "updated_at": "2025-06-05T01:42:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02197v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:06:43Z", "role": "primary", "sha256": "9048f54fbe64ee62117e37e3c9ad5b157fc70e83304d9ccb782f8782e487dd4a", "supports": ["environments", "scope", "notes"], "url": "https://arxiv.org/html/2503.02197"}], "id": "arxiv:2503.02197", "identifiers": {"arxiv": "2503.02197"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.02197"}], "name": "ATLaS: Agent Tuning via Learning Critical Steps", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The full paper was also inspected to confirm WebShop evaluation; the broader agent-learning method is classified as adjacent.", "published_at": "2025-03-04", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Selects critical steps from expert interaction trajectories for targeted agent fine-tuning, studying efficiency and generalization across tasks.", "topics": ["data-generation", "planning", "efficiency"], "updated_at": "2025-06-05", "url": "https://arxiv.org/abs/2503.02197"}
{"aliases": [], "authors": ["Wenjia Jiang", "Yangyang Zhuang", "Chenxi Song", "Xu Yang", "Joey Tianyi Zhou", "Chi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d396e4e249fc2aa1f4d53eadac38c3c814aae3f0208921b3388d5b009e6037b3", "submitted_at": "2025-03-04T04:34:09Z", "updated_at": "2025-04-15T02:32:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.02268v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.02268", "identifiers": {"arxiv": "2503.02268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.02268"}], "name": "AppAgentX: Evolving GUI Agents as Proficient Smartphone Users", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Records interaction history and converts repeated action sequences into reusable high-level shortcuts for mobile GUI tasks.", "topics": ["memory", "efficiency"], "updated_at": "2025-04-15", "url": "https://arxiv.org/abs/2503.02268"}
{"aliases": [], "authors": ["Jiahui Sun", "Zhichao Hua", "Yubin Xia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0c393d109ab0f8da2bf25bb39581236babd50f7554f2ec99679a99fb7215171", "submitted_at": "2025-03-04T08:44:30Z", "updated_at": "2025-09-24T05:11:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02403v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2503.02403", "identifiers": {"arxiv": "2503.02403"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.02403"}], "name": "AutoEval: A Practical Framework for Autonomous Evaluation of Mobile Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Generates reward signals from UI-state changes and uses a judging system to evaluate mobile agents with reduced manual evaluator construction.", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2025-09-24", "url": "https://arxiv.org/abs/2503.02403"}
{"aliases": [], "authors": ["Weimin Xiong", "Yifan Song", "Qingxiu Dong", "Bingchan Zhao", "Feifan Song", "Xun Wang", "Sujian Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3395107d8c7db3562bf75a3c22fee55d3da0564bd2f413491bc0d381fc601323", "submitted_at": "2025-03-04T14:54:45Z", "updated_at": "2025-09-10T16:45:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02682v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:06:43Z", "role": "primary", "sha256": "e634a3e576c94f1f7f5409d28a5c4c80c46141f1763c81337bb45f19793815f6", "supports": ["environments", "scope", "notes"], "url": "https://arxiv.org/html/2503.02682"}], "id": "arxiv:2503.02682", "identifiers": {"arxiv": "2503.02682"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.02682"}], "name": "MPO: Boosting LLM Agents with Meta Plan Optimization", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The full paper was also inspected to confirm WebShop evaluation; the broader agent-learning method is classified as adjacent.", "published_at": "2025-03-04", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Refines reusable high-level meta plans with execution feedback to guide language-agent planning across interactive environments.", "topics": ["planning", "memory"], "updated_at": "2025-09-10", "url": "https://arxiv.org/abs/2503.02682"}
{"aliases": [], "authors": ["Danqing Zhang", "Balaji Rama", "Jingyi Ni", "Shiying He", "Fu Zhao", "Kunyu Chen", "Arnold Chen", "Junyu Cao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df682f883bed5e0e9917074dfa130839fabaa318c55714f26943d944774854fe", "submitted_at": "2025-03-04T19:13:10Z", "updated_at": "2025-05-06T06:42:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02950v2"}, "publication": {"authors": ["Danqing Zhang", "Balaji Rama", "Jingyi Ni", "Shiying He", "Fu Zhao", "Kunyu Chen", "Arnold Chen", "Junyu Cao"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-demo.36", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-demo.36/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "449-455", "pdf_url": "https://aclanthology.org/2025.naacl-demo.36.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "LiteWebAgent: The Open-Source Suite for VLM-Based Web-Agent Applications", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-demo.36/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:58Z", "role": "primary", "sha256": "9b130228c049f324632f571dd9c9e74c9a293ee240351c7a6fd4fef3359afc12", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-demo.36/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.02950", "identifiers": {"arxiv": "2503.02950"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.02950"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-demo.36/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-demo.36.pdf"}], "name": "LiteWebAgent: The Open-Source Suite for VLM-Based Web-Agent Applications", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a modular visual web-agent stack with planning and memory components, exposed through a remote-browser application and Chrome extension.", "topics": ["web-navigation", "tool-integration", "planning"], "updated_at": "2025-05-06", "url": "https://arxiv.org/abs/2503.02950"}
{"aliases": [], "authors": ["Zhiyuan Huang", "Ziming Cheng", "Junting Pan", "Zhaohui Hou", "Mingjie Zhan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d45cacb33df18242eb3f47f6c89c56e41a146dd110e4b4dc5340062b9b981250", "submitted_at": "2025-03-05T05:30:22Z", "updated_at": "2025-04-16T15:34:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.03196v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.03196", "identifiers": {"arxiv": "2503.03196"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.03196"}], "name": "SpiritSight Agent: Advanced GUI Agent with One Look", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SpiritSight and GUI-Lasagne data, combining multilevel GUI annotations with visual block parsing for high-resolution interface navigation.", "topics": ["grounding", "screen-understanding", "data-generation"], "updated_at": "2025-04-16", "url": "https://arxiv.org/abs/2503.03196"}
{"aliases": [], "authors": ["Yuqi Zhou", "Shuai Wang", "Sunhao Dai", "Qinglin Jia", "Zhaocheng Du", "Zhenhua Dong", "Jun Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "380d00c46e5dd221b7c481a68b4fbdc00826e64e4b70a98f06a56f8c12b8f5e4", "submitted_at": "2025-03-05T18:56:16Z", "updated_at": "2025-03-05T18:56:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.03743v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2503.03743", "identifiers": {"arxiv": "2503.03743"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.03743"}], "name": "CHOP: Mobile Operating Assistant with Constrained High-frequency Optimized Subtask Planning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses human-planned subtasks to constrain and refine the intermediate plans connecting user goals with executable mobile operations.", "topics": ["mobile-navigation", "planning"], "updated_at": "2025-03-05", "url": "https://arxiv.org/abs/2503.03743"}
{"aliases": [], "authors": ["Zheng Hui", "Yinheng Li", "Dan zhao", "Tianyi Chen", "Colby Banbury", "Kazuhito Koishida"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc499e361f0bc46c3c5ee5a8abebd911396e383a004655de56db1ec7cec8e231", "submitted_at": "2025-01-27T08:29:17Z", "updated_at": "2025-01-27T08:29:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.04730v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.04730", "identifiers": {"arxiv": "2503.04730"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.04730"}], "name": "WinClick: GUI Grounding with Multimodal Large Language Models", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-01-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a screenshot-based Windows GUI agent and the WinSpot benchmark for evaluating localization of actionable interface regions.", "topics": ["grounding", "evaluation"], "updated_at": "2025-01-27", "url": "https://arxiv.org/abs/2503.04730"}
{"aliases": [], "authors": ["Ada Defne Tur", "Nicholas Meade", "Xing Han Lù", "Alejandra Zambrano", "Arkil Patel", "Esin Durmus", "Spandana Gella", "Karolina Stańczak", "Siva Reddy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4588022dd47624058db8a16d7f7763c591241298e94594b56bfeb4ab2e01737", "submitted_at": "2025-03-06T20:43:14Z", "updated_at": "2025-03-06T20:43:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.04957v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.04957", "identifiers": {"arxiv": "2503.04957"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.04957"}], "name": "SafeArena: Evaluating the Safety of Autonomous Web Agents", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-03-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Pairs benign and harmful web tasks to assess deliberate misuse risks and categorize agent behavior by risk level.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-03-06", "url": "https://arxiv.org/abs/2503.04957"}
{"aliases": [], "authors": ["Wenhao Wang", "Zijie Yu", "Rui Ye", "Jianqing Zhang", "Siheng Chen", "Yanfeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b408c850ae83e603a1b3d8303164b9f2d7be51580be99027275b0734ea8f56f2", "submitted_at": "2025-03-07T04:52:20Z", "updated_at": "2025-03-07T04:52:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.05143v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2503.05143", "identifiers": {"arxiv": "2503.05143"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.05143"}], "name": "FedMABench: Benchmarking Mobile Agents on Decentralized Heterogeneous User Data", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a benchmark for federated mobile-agent training under heterogeneous user and application distributions, combining datasets, algorithms, and base models.", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2025-03-07", "url": "https://arxiv.org/abs/2503.05143"}
{"aliases": [], "authors": ["Lexin Zhou", "Lorenzo Pacchiardi", "Fernando Martínez-Plumed", "Katherine M. Collins", "Yael Moros-Daval", "Seraphina Zhang", "Qinlin Zhao", "Yitian Huang", "Luning Sun", "Jonathan E. Prunty", "Zongqian Li", "Pablo Sánchez-García", "Kexin Jiang Chen", "Pablo A. M. Casares", "Jiyun Zu", "John Burden", "Behzad Mehrbakhsh", "David Stillwell", "Manuel Cebrian", "Jindong Wang", "Peter Henderson", "Sherry Tongshuang Wu", "Patrick C. Kyllonen", "Lucy Cheke", "Xing Xie", "José Hernández-Orallo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1163b1f7a4c9216cc549ab2e622fad2144c59c33b2936e99cf16d8db029481c1", "submitted_at": "2025-03-09T01:13:56Z", "updated_at": "2025-03-16T02:28:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.06378v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.06378", "identifiers": {"arxiv": "2503.06378"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.06378"}], "name": "General Scales Unlock AI Evaluation with Explanatory and Predictive Power", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-16", "url": "https://arxiv.org/abs/2503.06378"}
{"aliases": [], "authors": ["Fei Tang", "Yongliang Shen", "Hang Zhang", "Siqi Chen", "Guiyang Hou", "Wenqi Zhang", "Wenqiao Zhang", "Kaitao Song", "Weiming Lu", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebc104ee6a868092ae0215f84a3c5c71894bfb321671476a74b794e1645210c7", "submitted_at": "2025-03-09T06:14:17Z", "updated_at": "2025-03-09T06:14:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.06470v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.06470", "identifiers": {"arxiv": "2503.06470"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.06470"}], "name": "Think Twice, Click Once: Enhancing GUI Grounding via Fast and Slow Systems", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines rapid target prediction with deliberate interface analysis, adaptively selecting the grounding process according to task complexity.", "topics": ["grounding", "efficiency", "planning"], "updated_at": "2025-03-09", "url": "https://arxiv.org/abs/2503.06470"}
{"aliases": [], "authors": ["Yuxiang Zhang", "Yuqi Yang", "Jiangming Shu", "Xinyan Wen", "Jitao Sang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "696c49ce990b243184b20f14feb50f1c7c4a9169c17edeff8fde40572095115d", "submitted_at": "2025-03-09T12:19:47Z", "updated_at": "2025-03-09T12:19:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.06580v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2503.06580", "identifiers": {"arxiv": "2503.06580"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.06580"}], "name": "Agent models: Internalizing Chain-of-Action Generation into Reasoning models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-09", "url": "https://arxiv.org/abs/2503.06580"}
{"aliases": [], "authors": ["Fanqing Meng", "Lingxiao Du", "Zongkai Liu", "Zhixiang Zhou", "Quanfeng Lu", "Daocheng Fu", "Tiancheng Han", "Botian Shi", "Wenhai Wang", "Junjun He", "Kaipeng Zhang", "Ping Luo", "Yu Qiao", "Qiaosheng Zhang", "Wenqi Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7daf774ac3173f52c511171c9c807be9e24f602cb31b03c2700a68f6c1d50c7a", "submitted_at": "2025-03-10T14:23:12Z", "updated_at": "2025-04-15T14:22:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.07365v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2503.07365", "identifiers": {"arxiv": "2503.07365"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.07365"}], "name": "MM-Eureka: Exploring the Frontiers of Multimodal Reasoning with Rule-based Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-15", "url": "https://arxiv.org/abs/2503.07365"}
{"aliases": [], "authors": ["Konstantinos Vergopoulos", "Mark Niklas Müller", "Martin Vechev"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8f7998d9e9b10f7157129a544560cc38c45abe8c3422b260c1eb9f9573606076", "submitted_at": "2025-03-10T17:42:49Z", "updated_at": "2025-03-10T17:42:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.07701v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2503.07701", "identifiers": {"arxiv": "2503.07701"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.07701"}], "name": "Automated Benchmark Generation for Repository-Level Coding Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-10", "url": "https://arxiv.org/abs/2503.07701"}
{"aliases": [], "authors": ["Yixiao Song", "Katherine Thai", "Chau Minh Pham", "Yapei Chang", "Mazin Nadaf", "Mohit Iyyer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1cd6327559d11982fb4a97e739cb6b5234c509859692084ed78130241f6d252", "submitted_at": "2025-03-10T23:50:30Z", "updated_at": "2025-07-24T17:45:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.07919v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.07919", "identifiers": {"arxiv": "2503.07919"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.07919"}], "name": "BEARCUBS: A benchmark for computer-using web agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "BEARCUBS provides 111 live-web information-seeking questions with short answers and human-validated browsing trajectories. Its tasks require multimodal computer interaction, including video understanding and navigation of interactive 3D content.", "topics": ["evaluation", "web-navigation", "video"], "updated_at": "2025-07-24", "url": "https://arxiv.org/abs/2503.07919"}
{"aliases": [], "authors": ["Zhen Tan", "Jun Yan", "I-Hung Hsu", "Rujun Han", "Zifeng Wang", "Long T. Le", "Yiwen Song", "Yanfei Chen", "Hamid Palangi", "George Lee", "Anand Iyer", "Tianlong Chen", "Huan Liu", "Chen-Yu Lee", "Tomas Pfister"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bac2b66d2c03ab3eda3f7d5bcf83a9142874977e371acd685e72b176b3437050", "submitted_at": "2025-03-11T04:15:52Z", "updated_at": "2025-07-28T00:39:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.08026v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.08026", "identifiers": {"arxiv": "2503.08026"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.08026"}], "name": "In Prospect and Retrospect: Reflective Memory Management for Long-term Personalized Dialogue Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-28", "url": "https://arxiv.org/abs/2503.08026"}
{"aliases": [], "authors": ["Zitong Shi", "Guancheng Wan", "Wenke Huang", "Guibin Zhang", "Jiawei Shao", "Mang Ye", "Carl Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "783fb08cdaf1f53c337cc80d81e312cfb8b216972d3d0c90687e8d504b2168f7", "submitted_at": "2025-03-11T08:38:45Z", "updated_at": "2025-03-11T08:38:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.08175v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2503.08175", "identifiers": {"arxiv": "2503.08175"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.08175"}], "name": "Privacy-Enhancing Paradigms within Federated Multi-Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-11", "url": "https://arxiv.org/abs/2503.08175"}
{"aliases": [], "authors": ["Pei Yang", "Hai Ci", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d4423ee1e7047a6228f5b29684618fc4108af0694484a7ceddf6c3b53fbb06c", "submitted_at": "2025-03-12T10:38:15Z", "updated_at": "2025-03-12T10:38:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.09241v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2503.09241", "identifiers": {"arxiv": "2503.09241"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.09241"}], "name": "In-Context Defense in Computer Agents: An Empirical Study", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-03-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies in-context defenses for computer agents exposed to misleading pop-ups and other environmental instructions. It examines defensive reasoning before action planning and how a small set of examples changes resistance to these attacks.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-03-12", "url": "https://arxiv.org/abs/2503.09241"}
{"aliases": [], "authors": ["Di Zhao", "Longhui Ma", "Siwei Wang", "Miao Wang", "Zhao Lv"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3ea50d595c5d82b4a5d538b4b20725fa1f5afe5b29b03c022b726c307c8cca6e", "submitted_at": "2025-03-12T11:03:27Z", "updated_at": "2025-03-12T11:03:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.09263v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2503.09263", "identifiers": {"arxiv": "2503.09263"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.09263"}], "name": "COLA: A Scalable Multi-Agent Framework For Windows UI Task Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Coordinates specialized Windows UI agents with dynamic task scheduling, shared memory, and human-triggered rollback for recovering from errors.", "topics": ["planning", "memory", "human-in-the-loop"], "updated_at": "2025-03-12", "url": "https://arxiv.org/abs/2503.09263"}
{"aliases": [], "authors": ["Lutfi Eren Erdogan", "Nicholas Lee", "Sehoon Kim", "Suhong Moon", "Hiroki Furuta", "Gopala Anumanchipalli", "Kurt Keutzer", "Amir Gholami"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b03d84f29bb524d7950039624e0d460d8381f2ba63d70c69f5f958013730db9", "submitted_at": "2025-03-12T17:40:52Z", "updated_at": "2025-04-22T17:56:22Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.09572v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2503.09572", "identifiers": {"arxiv": "2503.09572"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.09572"}], "name": "Plan-and-Act: Improving Planning of Agents for Long-Horizon Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates planning from execution and creates synthetic plan annotations for training agents on extended web-navigation tasks.", "topics": ["planning", "long-horizon", "data-generation"], "updated_at": "2025-04-22", "url": "https://arxiv.org/abs/2503.09572"}
{"aliases": [], "authors": ["Arman Zharmagambetov", "Chuan Guo", "Ivan Evtimov", "Maya Pavlova", "Ruslan Salakhutdinov", "Kamalika Chaudhuri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bbb0726446c9679781b8e5930da0d3f406989a4b47759ce3440c9ad56d5839d8", "submitted_at": "2025-03-12T19:30:31Z", "updated_at": "2025-10-01T23:55:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.09780v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.09780", "identifiers": {"arxiv": "2503.09780"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.09780"}], "name": "AgentDAM: Privacy Leakage Evaluation for Autonomous Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates whether autonomous web agents minimize use of sensitive personal information when completing tasks that expose unnecessary private data.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-10-01", "url": "https://arxiv.org/abs/2503.09780"}
{"aliases": [], "authors": ["Advait Gupta", "NandaKiran Velaga", "Dang Nguyen", "Tianyi Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b77bbf2f746d7c78ebade6183e7206135e6bb9fe9b4b2853991982177454b60c", "submitted_at": "2025-03-13T17:55:45Z", "updated_at": "2025-03-13T17:55:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.10613v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.10613", "identifiers": {"arxiv": "2503.10613"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.10613"}], "name": "CoSTA$\\ast$: Cost-Sensitive Toolpath Agent for Multi-turn Image Editing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-13", "url": "https://arxiv.org/abs/2503.10613"}
{"aliases": [], "authors": ["Lukas Aichberger", "Alasdair Paren", "Guohao Li", "Philip Torr", "Yarin Gal", "Adel Bibi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e5622a83d74c18761a7ab88c98682d3903ce0b6700c7e9814f069bfba26e796", "submitted_at": "2025-03-13T18:59:12Z", "updated_at": "2026-01-27T18:10:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.10809v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.10809", "identifiers": {"arxiv": "2503.10809"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.10809"}], "name": "MIP against Agent: Malicious Image Patches Hijacking Multimodal OS Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-27", "url": "https://arxiv.org/abs/2503.10809"}
{"aliases": [], "authors": ["Chaoyun Zhang", "Shilin He", "Liqun Li", "Si Qin", "Yu Kang", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43a533866496d3ebd3e34c04142a702e4a3aaa387c3830e3bb993fdbb0ce8432", "submitted_at": "2025-03-14T04:26:21Z", "updated_at": "2025-06-23T13:01:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.11069v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.11069", "identifiers": {"arxiv": "2503.11069"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.11069"}], "name": "API Agents vs. GUI Agents: Divergence and Convergence", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares API-driven and GUI-driven agents across architecture, development, and interaction considerations, discussing when hybrid approaches are useful.", "topics": ["tool-integration", "survey", "automation"], "updated_at": "2025-06-23", "url": "https://arxiv.org/abs/2503.11069"}
{"aliases": [], "authors": ["Yibin Xu", "Liang Yang", "Hao Chen", "Hua Wang", "Zhi Chen", "Yaohua Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2e6250e87ea9340914b457c4b0d471450f8cf49d2e2d53abd8e9e76105e908a", "submitted_at": "2025-03-14T08:16:02Z", "updated_at": "2025-03-14T08:16:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.11170v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.11170", "identifiers": {"arxiv": "2503.11170"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.11170"}], "name": "DeskVision: Large Scale Desktop Region Captioning for Advanced GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AutoCaptioner, DeskVision, and GUIExplorer for generating desktop region descriptions and training interface-understanding models.", "topics": ["data-generation", "grounding", "screen-understanding"], "updated_at": "2025-03-14", "url": "https://arxiv.org/abs/2503.11170"}
{"aliases": [], "authors": ["Weiming Ren", "Wentao Ma", "Huan Yang", "Cong Wei", "Ge Zhang", "Wenhu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d09f8d799a684b4087731df0e82251f58a3d8d555cd08a800bcb718c2d583936", "submitted_at": "2025-03-14T16:45:23Z", "updated_at": "2025-07-16T11:39:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.11579v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.11579", "identifiers": {"arxiv": "2503.11579"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.11579"}], "name": "Vamba: Understanding Hour-Long Videos with Hybrid Mamba-Transformers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-16", "url": "https://arxiv.org/abs/2503.11579"}
{"aliases": [], "authors": ["Yihong Jin", "Ze Yang", "Xinhe Xu", "Yihan Zhang", "Shuyang Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f2b7436cb190a5f53069243d57e7a9e264d9d3ea767593d7b732c30b8dcd36ab", "submitted_at": "2025-03-15T18:45:33Z", "updated_at": "2025-03-15T18:45:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.12228v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.12228", "identifiers": {"arxiv": "2503.12228"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.12228"}], "name": "Adaptive Fault Tolerance Mechanisms of Large Language Models in Cloud Computing Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-15", "url": "https://arxiv.org/abs/2503.12228"}
{"aliases": [], "authors": ["Fanhu Zeng", "Hao Tang", "Yihua Shao", "Siyu Chen", "Ling Shao", "Yan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4878a982a9e7fb5dcd5d2490cac07250d3fc6724a70807fa4b1db0231b34529", "submitted_at": "2025-03-16T11:32:34Z", "updated_at": "2025-08-22T05:01:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.12461v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.12461", "identifiers": {"arxiv": "2503.12461"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.12461"}], "name": "MambaIC: State Space Models for High-Performance Learned Image Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-22", "url": "https://arxiv.org/abs/2503.12461"}
{"aliases": [], "authors": ["Fanbin Lu", "Zhisheng Zhong", "Ziqin Wei", "Shu Liu", "Chi-Wing Fu", "Jiaya Jia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e883289bafee01759a0f4f3f7051cbbebdcd8059e7520d68c25ba313b8da69a6", "submitted_at": "2025-03-16T14:53:43Z", "updated_at": "2025-03-24T16:33:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.12532v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.12532", "identifiers": {"arxiv": "2503.12532"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.12532"}], "name": "STEVE: A Step Verification Pipeline for Computer-use Agent Training", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Labels individual computer-use actions by comparing surrounding screenshots, then trains agents from both successful and unsuccessful steps.", "topics": ["data-generation", "evaluation", "reinforcement-learning"], "updated_at": "2025-03-24", "url": "https://arxiv.org/abs/2503.12532"}
{"aliases": [], "authors": ["Naveen Krishnan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c7d22ba984bb2ddf84a45dd9abfbfc6f770682133a124229ae680f2d6afbed3", "submitted_at": "2025-03-16T23:07:48Z", "updated_at": "2025-03-16T23:07:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.12687v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2503.12687", "identifiers": {"arxiv": "2503.12687"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.12687"}], "name": "AI Agents: Evolution, Architecture, and Real-World Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-16", "url": "https://arxiv.org/abs/2503.12687"}
{"aliases": [], "authors": ["Trisanth Srinivasan", "Santosh Patapati"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ad657a8e90b268ba60790caf8c6d9f872a86efca8852e1408d40a8583bf96061", "submitted_at": "2025-03-18T02:33:27Z", "updated_at": "2025-07-06T08:41:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.13843v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.13843", "identifiers": {"arxiv": "2503.13843"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.13843"}], "name": "WebNav: An Intelligent Agent for Voice-Controlled Web Navigation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-06", "url": "https://arxiv.org/abs/2503.13843"}
{"aliases": [], "authors": ["Ziwei Wang", "Weizhi Chen", "Leyang Yang", "Sheng Zhou", "Shengchu Zhao", "Hanbei Zhan", "Jiongchao Jin", "Liangcheng Li", "Zirui Shao", "Jiajun Bu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b260fd1d7e28c8263710f116638e7aa7b51f46aac15199c8a29092161a13f37d", "submitted_at": "2025-03-18T08:32:22Z", "updated_at": "2025-03-18T08:32:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.14021v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2503.14021", "identifiers": {"arxiv": "2503.14021"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.14021"}], "name": "MP-GUI: Modality Perception with MLLMs for GUI Understanding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses specialized graphical, textual, and spatial perception components with a data-collection pipeline to improve GUI understanding.", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2025-03-18", "url": "https://arxiv.org/abs/2503.14021"}
{"aliases": [], "authors": ["Jiacheng Guo", "Yue Wu", "Jiahao Qiu", "Kaixuan Huang", "Xinzhe Juan", "Ling Yang", "Mengdi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cad5256bff4b5ecb346b266b7d29c9c86bba73905eef4769e82666ef4964b40a", "submitted_at": "2025-03-18T17:58:28Z", "updated_at": "2025-03-18T17:58:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.14495v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.14495", "identifiers": {"arxiv": "2503.14495"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.14495"}], "name": "Temporal Consistency for LLM Reasoning Process Error Identification", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-18", "url": "https://arxiv.org/abs/2503.14495"}
{"aliases": [], "authors": ["Wei Tang", "Yanpeng Sun", "Qinying Gu", "Zechao Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a157c5a6ef865aef40580e43806abb55ba64d12ccdc21e7b3b9919cb528400bf", "submitted_at": "2025-03-19T17:08:13Z", "updated_at": "2025-07-16T13:53:05Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2503.15426v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.15426", "identifiers": {"arxiv": "2503.15426"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.15426"}], "name": "Visual Position Prompt for MLLM based Visual Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-16", "url": "https://arxiv.org/abs/2503.15426"}
{"aliases": [], "authors": ["Piero A. Bonatti", "John Domingue", "Anna Lisa Gentile", "Andreas Harth", "Olaf Hartig", "Aidan Hogan", "Katja Hose", "Ernesto Jimenez-Ruiz", "Deborah L. McGuinness", "Chang Sun", "Ruben Verborgh", "Jesse Wright"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c778da33baa88916fbcb3e02a33f760dccea8b193d25087df9b56cd42c3c248", "submitted_at": "2025-01-31T12:26:27Z", "updated_at": "2025-01-31T12:26:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.15515v1"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.15515", "identifiers": {"arxiv": "2503.15515"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.15515"}], "name": "Towards Computer-Using Personal Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-01-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes computer-using personal agents connected to user-controlled data repositories, discussing privacy, interoperability, and collaborative tasks.", "topics": ["safety-security", "memory", "human-in-the-loop"], "updated_at": "2025-01-31", "url": "https://arxiv.org/abs/2503.15515"}
{"aliases": [], "authors": ["Shravan Nayak", "Xiangru Jian", "Kevin Qinghong Lin", "Juan A. Rodriguez", "Montek Kalsi", "Rabiul Awal", "Nicolas Chapados", "M. Tamer Özsu", "Aishwarya Agrawal", "David Vazquez", "Christopher Pal", "Perouz Taslakian", "Spandana Gella", "Sai Rajeswar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5abf3c7b08a4331d4f634d8cc63a7f8ff40713f7344231887dac1b5e061a8eae", "submitted_at": "2025-03-19T19:26:17Z", "updated_at": "2025-05-06T17:43:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.15661v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "arxiv:2503.15661", "identifiers": {"arxiv": "2503.15661"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.15661"}], "name": "UI-Vision: A Desktop-centric GUI Benchmark for Visual Perception and Interaction", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-03-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Provides annotated desktop demonstrations for evaluating element grounding, layout grounding, and action prediction.", "topics": ["grounding", "evaluation", "screen-understanding"], "updated_at": "2025-05-06", "url": "https://arxiv.org/abs/2503.15661"}
{"aliases": [], "authors": ["Gaole Dai", "Shiqi Jiang", "Ting Cao", "Yuanchun Li", "Yuqing Yang", "Rui Tan", "Mo Li", "Lili Qiu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6449576cb8ba45569b57844d87f22d01645b3639a222c648d914dc38e3421553", "submitted_at": "2025-03-20T08:25:00Z", "updated_at": "2026-02-21T01:36:43Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2503.15937v5"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.15937", "identifiers": {"arxiv": "2503.15937"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.15937"}], "name": "Advancing Mobile GUI Agents: A Verifier-Driven Approach to Practical Deployment", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces V-Droid, which verifies candidate actions in a discrete action space rather than directly generating each mobile operation.", "topics": ["mobile-navigation", "efficiency", "planning"], "updated_at": "2026-02-21", "url": "https://arxiv.org/abs/2503.15937"}
{"aliases": [], "authors": ["Zhihang Liu", "Chen-Wei Xie", "Pandeng Li", "Liming Zhao", "Longxiang Tang", "Yun Zheng", "Chuanbin Liu", "Hongtao Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f3235c0ec3b16c1f16ca1233cd9b723a9b06c25c9a9199a635b874371b1d7e6", "submitted_at": "2025-03-20T11:09:18Z", "updated_at": "2025-03-20T11:09:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.16036v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.16036", "identifiers": {"arxiv": "2503.16036"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.16036"}], "name": "Hybrid-Level Instruction Injection for Video Token Compression in Multi-modal Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-20", "url": "https://arxiv.org/abs/2503.16036"}
{"aliases": [], "authors": ["Atharv Singh Patlan", "Peiyao Sheng", "S. Ashwin Hebbar", "Prateek Mittal", "Pramod Viswanath"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cbb4844de6aac9118521e4348d1df7af35147b688d6b8718a40e2c9deb1f294c", "submitted_at": "2025-03-20T15:44:31Z", "updated_at": "2025-07-09T01:38:20Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.16248v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.16248", "identifiers": {"arxiv": "2503.16248"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.16248"}], "name": "Real AI Agents with Fake Memories: Fatal Context Manipulation Attacks on Web3 Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-09", "url": "https://arxiv.org/abs/2503.16248"}
{"aliases": [], "authors": ["Pengzhou Cheng", "Zheng Wu", "Zongru Wu", "Aston Zhang", "Zhuosheng Zhang", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa6201528b13c6a1aae8dcd745f4f5d995e79982ef7c54a1962e2f9035233293", "submitted_at": "2025-02-26T12:31:16Z", "updated_at": "2025-07-14T15:56:44Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.16465v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.16465", "identifiers": {"arxiv": "2503.16465"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.16465"}], "name": "OS-Kairos: Adaptive Interaction for MLLM-Powered GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-02-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses step-level confidence estimates to decide when a GUI agent should act autonomously or request human intervention in complex situations.", "topics": ["human-in-the-loop", "safety-security", "mobile-navigation"], "updated_at": "2025-07-14", "url": "https://arxiv.org/abs/2503.16465"}
{"aliases": [], "authors": ["Li Zhang", "Longxi Gao", "Mengwei Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e450216f8f802d8f3d16c5886097cef801d051951af0c57fdf195cc176ac744f", "submitted_at": "2025-03-21T01:52:43Z", "updated_at": "2025-03-21T01:52:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.16788v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.16788", "identifiers": {"arxiv": "2503.16788"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.16788"}], "name": "Does Chain-of-Thought Reasoning Help Mobile GUI Agent? An Empirical Study", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-03-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study compares matched reasoning and non-reasoning vision-language models on static mobile GUI tasks and interactive AndroidWorld navigation. It finds that chain-of-thought effects vary across tasks and model families.", "topics": ["evaluation", "mobile-navigation", "planning"], "updated_at": "2025-03-21", "url": "https://arxiv.org/abs/2503.16788"}
{"aliases": [], "authors": ["Xiaofeng Mao", "Yuefeng Chen", "Rong Zhang", "Hui Xue", "Zhao Li", "Hang Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e01918698fc9e3e2715110ca46b556b468a8c464d49bc173076b6f2a3c9988a", "submitted_at": "2025-03-21T09:43:42Z", "updated_at": "2025-03-21T09:43:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.16975v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.16975", "identifiers": {"arxiv": "2503.16975"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.16975"}], "name": "EasyRobust: A Comprehensive and Easy-to-use Toolkit for Robust and Generalized Vision", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-21", "url": "https://arxiv.org/abs/2503.16975"}
{"aliases": [], "authors": ["Oucheng Huang", "Yuhang Ma", "Zeng Zhao", "Mingrui Wu", "Jiayi Ji", "Rongsheng Zhang", "Zhipeng Hu", "Xiaoshuai Sun", "Rongrong Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75b0e0c817aa3f31be0f98c1f95a2b7b167afd66d0f59577d5fa792d3da86f22", "submitted_at": "2025-03-22T06:48:50Z", "updated_at": "2025-09-17T06:29:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.17671v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.17671", "identifiers": {"arxiv": "2503.17671"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.17671"}], "name": "ComfyGPT: A Self-Optimizing Multi-Agent System for Comprehensive ComfyUI Workflow Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-17", "url": "https://arxiv.org/abs/2503.17671"}
{"aliases": [], "authors": ["Yuchen Sun", "Shanhui Zhao", "Tao Yu", "Hao Wen", "Samith Va", "Mengwei Xu", "Yuanchun Li", "Chongyang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "80bb507bf3cdccf6f023d0c334899d9d7099c362e9f37c314b15ff9388bae65c", "submitted_at": "2025-03-22T09:30:37Z", "updated_at": "2025-03-22T09:30:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.17709v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.17709", "identifiers": {"arxiv": "2503.17709"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.17709"}], "name": "GUI-Xplore: Empowering Generalizable GUI Agents with One Exploration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI-Xplore exploration videos and downstream tasks, with an agent that combines action-aware modeling and graph-based environment reasoning.", "topics": ["data-generation", "planning", "video"], "updated_at": "2025-03-22", "url": "https://arxiv.org/abs/2503.17709"}
{"aliases": [], "authors": ["Samuel Schmidgall", "Michael Moor"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "704bb440db680617d83c9c1dac9bcfe9d554c26779e69b9ccb4fec8dc7281def", "submitted_at": "2025-03-23T15:16:42Z", "updated_at": "2025-03-23T15:16:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.18102v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.18102", "identifiers": {"arxiv": "2503.18102"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.18102"}], "name": "AgentRxiv: Towards Collaborative Autonomous Research", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-23", "url": "https://arxiv.org/abs/2503.18102"}
{"aliases": [], "authors": ["Jungjae Lee", "Dongjae Lee", "Chihun Choi", "Youngmin Im", "Jaeyoung Wi", "Kihong Heo", "Sangeun Oh", "Sunjae Lee", "Insik Shin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "10e7a9a4471555c6e4001cd10c1e4b375b503adfe6ea48d5808d1d7721dfc162", "submitted_at": "2025-03-24T09:46:05Z", "updated_at": "2025-09-11T04:15:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.18492v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.18492", "identifiers": {"arxiv": "2503.18492"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.18492"}], "name": "VeriSafe Agent: Safeguarding Mobile GUI Agent via Logic-based Action Verification", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Translates user instructions into formal specifications and applies runtime logic-based checks to proposed mobile GUI actions.", "topics": ["safety-security", "mobile-navigation"], "updated_at": "2025-09-11", "url": "https://arxiv.org/abs/2503.18492"}
{"aliases": [], "authors": ["Noam Kahlon", "Guy Rom", "Anatoly Efros", "Filippo Galgani", "Omri Berkovitch", "Sapir Caduri", "William E. Bishop", "Oriana Riva", "Ido Dagan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a49d662eddf84690395b025e50f3378c3be2a3fc8abe02bd7202c6ff879bbe3", "submitted_at": "2025-03-25T10:46:08Z", "updated_at": "2025-03-25T10:46:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.19537v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2503.19537", "identifiers": {"arxiv": "2503.19537"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.19537"}], "name": "Agent-Initiated Interaction in Phone UI Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Defines when phone agents should request user interaction and introduces AndroidInteraction annotations for deciding whether and how to engage the user.", "topics": ["human-in-the-loop", "mobile-navigation", "evaluation"], "updated_at": "2025-03-25", "url": "https://arxiv.org/abs/2503.19537"}
{"aliases": [], "authors": ["Sejin Lee", "Jian Kim", "Haon Park", "Ashkan Yousefpour", "Sangyoon Yu", "Min Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e90cbc547703a2673cfb30bbe9079496ce4c076b51f39e0557f5db9048947c8", "submitted_at": "2025-03-26T07:08:15Z", "updated_at": "2025-06-08T05:39:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.20279v3"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2503.20279", "identifiers": {"arxiv": "2503.20279"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.20279"}], "name": "sudo rm -rf agentic_security", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-03-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates security failures of computer-use agents under adversarial requests and screen-context manipulation. The study examines refusal behavior while agents perform tasks in desktop and web environments.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-06-08", "url": "https://arxiv.org/abs/2503.20279"}
{"aliases": [], "authors": ["Yiqiao Jin", "Stefano Petrangeli", "Yu Shen", "Gang Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0888d0da1c5da9cf9956fd8b145a09f463aa432a60b142c4aea8d20b4aa84c4d", "submitted_at": "2025-03-26T20:41:24Z", "updated_at": "2025-03-26T20:41:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.20978v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2503.20978", "identifiers": {"arxiv": "2503.20978"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.20978"}], "name": "ScreenLLM: Stateful Screen Schema for Efficient Action Understanding and Prediction", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Represents interface interactions with a stateful screen schema to model user behavior and support action understanding and prediction.", "topics": ["screen-understanding", "memory"], "updated_at": "2025-03-26", "url": "https://arxiv.org/abs/2503.20978"}
{"aliases": [], "authors": ["Zhengxi Lu", "Yuxiang Chai", "Yaxuan Guo", "Xi Yin", "Liang Liu", "Hao Wang", "Han Xiao", "Shuai Ren", "Guanjing Xiong", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b794a2f59cbe26a889b242ff47a7cd982f6367f759ddabcf99d96f31e5c122da", "submitted_at": "2025-03-27T15:39:30Z", "updated_at": "2025-05-24T08:46:08Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2503.21620v5"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.21620", "identifiers": {"arxiv": "2503.21620"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.21620"}], "name": "UI-R1: Enhancing Efficient Action Prediction of GUI Agents by Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses rule-based action rewards and policy optimization to train GUI action prediction from a compact set of mobile tasks.", "topics": ["reinforcement-learning", "grounding"], "updated_at": "2025-05-24", "url": "https://arxiv.org/abs/2503.21620"}
{"aliases": [], "authors": ["Patrice Bechard", "Chao Wang", "Amirhossein Abaskohi", "Juan Rodriguez", "Christopher Pal", "David Vazquez", "Spandana Gella", "Sai Rajeswar", "Perouz Taslakian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "59eefcf29997b2ed674265cae019ecc3c2a9c844be870574204748fd7fbeaea3", "submitted_at": "2025-03-27T18:04:05Z", "updated_at": "2026-01-12T18:27:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.21889v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.21889", "identifiers": {"arxiv": "2503.21889"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.21889"}], "name": "StarFlow: Generating Structured Workflow Outputs From Sketch Images", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2503.21889"}
{"aliases": [], "authors": ["Rati Devidze"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "389894e2e529598cca74d19fdc1393ca2e8459d4184cbbc1de61aec854a38628", "submitted_at": "2025-03-27T19:48:02Z", "updated_at": "2025-03-27T19:48:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.21949v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.21949", "identifiers": {"arxiv": "2503.21949"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.21949"}], "name": "Reward Design for Reinforcement Learning Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-03-27", "url": "https://arxiv.org/abs/2503.21949"}
{"aliases": [], "authors": ["Zhaorun Chen", "Mintong Kang", "Bo Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "362fe117797181225641b71b23df95d82c2f90ebf2bbfa1bebea56427e502091", "submitted_at": "2025-03-26T17:58:40Z", "updated_at": "2025-11-27T22:01:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.22738v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2503.22738", "identifiers": {"arxiv": "2503.22738"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.22738"}], "name": "ShieldAgent: Shielding Agents via Verifiable Safety Policy Reasoning", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-03-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes a guardrail agent that reasons over explicit safety policies and checks action trajectories, with a web-agent safety benchmark.", "topics": ["safety-security"], "updated_at": "2025-11-27", "url": "https://arxiv.org/abs/2503.22738"}
{"aliases": [], "authors": ["Liangbo Ning", "Ziran Liang", "Zhuohang Jiang", "Haohao Qu", "Yujuan Ding", "Wenqi Fan", "Xiao-yong Wei", "Shanru Lin", "Hui Liu", "Philip S. Yu", "Qing Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4b214bd8e5a34c0e8ca16a3be4fcf638a59a17ca524276dbef3c6b75f08398bb", "submitted_at": "2025-03-30T08:15:44Z", "updated_at": "2025-08-05T11:47:41Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2503.23350v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2503.23350", "identifiers": {"arxiv": "2503.23350"}, "kind": "survey", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.23350"}], "name": "A Survey of WebAgents: Towards Next-Generation AI Agents for Web Automation with Large Foundation Models", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-03-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Surveys foundation-model web agents through their architectures, training methods and trustworthiness. It organizes research on completing browser tasks from user instructions and identifies directions for further development.", "topics": ["survey", "web-navigation", "safety-security"], "updated_at": "2025-08-05", "url": "https://arxiv.org/abs/2503.23350"}
{"aliases": [], "authors": ["Yucheng Shi", "Wenhao Yu", "Jingyuan Huang", "Wenlin Yao", "Wenhu Chen", "Ninghao Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a6811c6d0c3b30d4e6af2f7b0beef8aa8f1ad99ceb325182305e571fe5c398f", "submitted_at": "2025-03-30T13:26:00Z", "updated_at": "2026-02-24T05:24:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.23434v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2503.23434", "identifiers": {"arxiv": "2503.23434"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.23434"}], "name": "Towards Trustworthy GUI Agents: A Survey", "notes": "", "published_at": "2025-03-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Frames GUI-agent trustworthiness as an execution gap and reviews failures, attacks, and defenses across perception, reasoning, and interaction.", "topics": ["survey", "safety-security"], "updated_at": "2026-02-24", "url": "https://arxiv.org/abs/2503.23434"}
{"aliases": [], "authors": ["Yingwei Ma", "Yongbin Li", "Yihong Dong", "Xue Jiang", "Rongyu Cao", "Jue Chen", "Fei Huang", "Binhua Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "579500087cb4dbd64580d983af737b9f559c8d693c24442aaf00b659051b8fc9", "submitted_at": "2025-03-31T07:31:32Z", "updated_at": "2025-04-08T12:36:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.23803v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2503.23803", "identifiers": {"arxiv": "2503.23803"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2503.23803"}], "name": "Thinking Longer, Not Larger: Enhancing Software Engineering Agents via Scaling Test-Time Compute", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-03-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-08", "url": "https://arxiv.org/abs/2503.23803"}
{"aliases": [], "authors": ["Ziming Cheng", "Zhiyuan Huang", "Junting Pan", "Zhaohui Hou", "Mingjie Zhan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bb09b54379d2fbc5b489f24a27f6311b91bb0821eefe5187aa160ffb00639d9", "submitted_at": "2025-03-31T14:56:24Z", "updated_at": "2025-08-24T08:39:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.24180v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2503.24180", "identifiers": {"arxiv": "2503.24180"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.24180"}], "name": "Navi-plus: Managing Ambiguous GUI Navigation Tasks with Follow-up Questions", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-03-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Navi-plus evaluates GUI agents that ask follow-up questions to complete ambiguous user instructions. It provides GUI clarification question-answer pairs and Dual-Stream Trajectory Evaluation for this interactive setting.", "topics": ["human-in-the-loop", "evaluation", "planning"], "updated_at": "2025-08-24", "url": "https://arxiv.org/abs/2503.24180"}
{"aliases": [], "authors": ["Saaket Agashe", "Kyle Wong", "Vincent Tu", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "11958f69503db514a91af8721870a4cbc1fa19dabf94e3f4ce4118d2fad9862a", "submitted_at": "2025-04-01T15:40:27Z", "updated_at": "2025-04-01T15:40:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.00906v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.00906", "identifiers": {"arxiv": "2504.00906"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.00906"}], "name": "Agent S2: A Compositional Generalist-Specialist Framework for Computer Use Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-04-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines generalist and specialist models with hierarchical planning and multiple grounding methods for computer-use tasks.", "topics": ["grounding", "planning"], "updated_at": "2025-04-01", "url": "https://arxiv.org/abs/2504.00906"}
{"aliases": [], "authors": ["Min Shi", "Shihao Wang", "Chieh-Yun Chen", "Jitesh Jain", "Kai Wang", "Junjun Xiong", "Guilin Liu", "Zhiding Yu", "Humphrey Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87abcedbc3693f958b95f76c69af829743c0bec03b64a4e52153cfde62a3f527", "submitted_at": "2025-04-02T03:24:58Z", "updated_at": "2025-04-02T03:24:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.01328v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2504.01328", "identifiers": {"arxiv": "2504.01328"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.01328"}], "name": "Slow-Fast Architecture for Video Multi-Modal Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-02", "url": "https://arxiv.org/abs/2504.01328"}
{"aliases": [], "authors": ["Hanxi Fang", "Supawit Chockchowwat", "Hari Sundaram", "Yongjoo Park"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "450f3f3fb0fa31a44b8806678c84bfcd821564973f5af8d50c5b7090a3cf2206", "submitted_at": "2025-04-02T05:41:12Z", "updated_at": "2025-04-02T05:41:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.01377v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}], "id": "arxiv:2504.01377", "identifiers": {"arxiv": "2504.01377"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.01377"}], "name": "Large-scale Evaluation of Notebook Checkpointing with AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-02", "url": "https://arxiv.org/abs/2504.01377"}
{"aliases": [], "authors": ["Tianci Xue", "Weijian Qi", "Tianneng Shi", "Chan Hee Song", "Boyu Gou", "Dawn Song", "Huan Sun", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccad667f7c5bb75e58d75d5eab5cc24c7c3610be9eed1a09079de5e23c85af04", "submitted_at": "2025-04-02T05:51:29Z", "updated_at": "2025-10-08T02:22:45Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.01382v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.01382", "identifiers": {"arxiv": "2504.01382"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.01382"}], "name": "An Illusion of Progress? Assessing the Current State of Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Online-Mind2Web, with 300 tasks across 136 live websites, and examines how benchmark design affects assessments of web agents. It also develops an LLM judge for task completion and compares its judgments with human evaluation.", "topics": ["evaluation", "web-navigation"], "updated_at": "2025-10-08", "url": "https://arxiv.org/abs/2504.01382"}
{"aliases": [], "authors": ["Antoine Chevrot", "Alexandre Vernotte", "Jean-Rémy Falleri", "Xavier Blanc", "Bruno Legeard"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76d0ffb55478570499c8cddf92a27d346cf11749f8ae5f8e128c87653b2bdbdc", "submitted_at": "2025-04-02T08:48:01Z", "updated_at": "2025-04-02T08:48:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.01495v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.01495", "identifiers": {"arxiv": "2504.01495"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.01495"}], "name": "Are Autonomous Web Agents Good Testers?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adapts web agents to execute natural-language test steps, verify assertions, and issue test verdicts. The study introduces 113 passing and failing test cases across three offline web applications and compares two agent implementations with qualitative failure analysis.", "topics": ["automation", "evaluation", "web-navigation"], "updated_at": "2025-04-02", "url": "https://arxiv.org/abs/2504.01495"}
{"aliases": [], "authors": ["Xiaolei Li", "Jialun Cao", "Yepang Liu", "Shing-Chi Cheung", "Hailong Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a646bb2136cdaf47a0d8b3f485d3e1973fc15ac4f844e1beddedf6021b56ffe4", "submitted_at": "2025-04-03T07:45:09Z", "updated_at": "2025-04-03T07:45:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.02357v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:12:09Z", "role": "primary", "sha256": "054a5a305d86c7fb9107d467d194fc4a64aa4f5af23d214bea7f52c2f2c9fd35", "supports": ["environments"], "url": "https://arxiv.org/html/2504.02357"}], "id": "arxiv:2504.02357", "identifiers": {"arxiv": "2504.02357"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.02357"}], "name": "ReuseDroid: A VLM-empowered Android UI Test Migrator Boosted by Active Feedback", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The platform or GUI evaluation setting was additionally checked in the primary full text.", "published_at": "2025-04-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "ReuseDroid migrates Android GUI tests between functionally similar apps using VLM agents that separate shared test intent from app-specific interaction flows. It introduces LinPro, a dataset of 578 migration tasks across 39 apps, to evaluate test transfer.", "topics": ["automation", "evaluation", "cross-app"], "updated_at": "2025-04-03", "url": "https://arxiv.org/abs/2504.02357"}
{"aliases": [], "authors": ["Yanming Wan", "Jiaxing Wu", "Marwa Abdulhai", "Lior Shani", "Natasha Jaques"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aff3b863c6c0543cd7aebd895cf9173173b154c1c4d31ba8dd4d3a9cd509dcea", "submitted_at": "2025-04-04T06:35:02Z", "updated_at": "2025-10-02T16:36:08Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.03206v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.03206", "identifiers": {"arxiv": "2504.03206"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.03206"}], "name": "Enhancing Personalized Multi-Turn Dialogue with Curiosity Reward", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-02", "url": "https://arxiv.org/abs/2504.03206"}
{"aliases": [], "authors": ["Akshara Prabhakar", "Zuxin Liu", "Ming Zhu", "Jianguo Zhang", "Tulika Awalgaonkar", "Shiyu Wang", "Zhiwei Liu", "Haolin Chen", "Thai Hoang", "Juan Carlos Niebles", "Shelby Heinecke", "Weiran Yao", "Huan Wang", "Silvio Savarese", "Caiming Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "17d4b2f30011c83cafb4a04ebcc9d37c5c68c2e32f6b3214c6e143980cef6d4a", "submitted_at": "2025-04-04T17:13:57Z", "updated_at": "2025-07-19T17:39:17Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.03601v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.03601", "identifiers": {"arxiv": "2504.03601"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.03601"}], "name": "APIGen-MT: Agentic Pipeline for Multi-Turn Data Generation via Simulated Agent-Human Interplay", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-19", "url": "https://arxiv.org/abs/2504.03601"}
{"aliases": [], "authors": ["Haoren Zhao", "Tianyi Chen", "Zhen Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ededfa864150f1b7b4ba53b265b5cbaa60bcf42690a8e22f5c13f915b6cd0108", "submitted_at": "2025-04-07T03:58:45Z", "updated_at": "2025-04-07T03:58:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.04716v1"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.04716", "identifiers": {"arxiv": "2504.04716"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.04716"}], "name": "On the Robustness of GUI Grounding Models Against Image Attacks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates GUI grounding models under natural image noise, untargeted adversarial perturbations, and targeted image attacks. Experiments cover mobile, desktop, and web interfaces and analyze sensitivity to image resolution and adversarial changes.", "topics": ["grounding", "robustness", "safety-security", "evaluation"], "updated_at": "2025-04-07", "url": "https://arxiv.org/abs/2504.04716"}
{"aliases": [], "authors": ["Zora Zhiruo Wang", "Apurva Gandhi", "Graham Neubig", "Daniel Fried"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54688f083a219d1f6b871967cecb7319dfe17c3eb69f6a25affe25a53149fae4", "submitted_at": "2025-04-09T12:25:37Z", "updated_at": "2025-08-29T15:46:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.06821v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.06821", "identifiers": {"arxiv": "2504.06821"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.06821"}], "name": "Inducing Programmatic Skills for Agentic Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-29", "url": "https://arxiv.org/abs/2504.06821"}
{"aliases": [], "authors": ["Boyuan Zheng", "Michael Y. Fatemi", "Xiaolong Jin", "Zora Zhiruo Wang", "Apurva Gandhi", "Yueqi Song", "Yu Gu", "Jayanth Srinivasa", "Gaowen Liu", "Graham Neubig", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f0489433a41b49ab8ab42eea57503685ba60beadabea7822134e96d865a7435", "submitted_at": "2025-04-09T17:51:50Z", "updated_at": "2025-04-09T17:51:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.07079v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.07079", "identifiers": {"arxiv": "2504.07079"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.07079"}], "name": "SkillWeaver: Web Agents can Self-Improve by Discovering and Honing Skills", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "SkillWeaver explores websites, practices interactions, and converts the resulting experience into a reusable library of callable skill APIs. Its WebArena and live-website experiments examine skill refinement, composition, and transfer from stronger agents to weaker ones.", "topics": ["web-navigation", "tool-integration", "planning"], "updated_at": "2025-04-09", "url": "https://arxiv.org/abs/2504.07079"}
{"aliases": [], "authors": ["Kaixin Li", "Ziyang Meng", "Hongzhan Lin", "Ziyang Luo", "Yuchen Tian", "Jing Ma", "Zhiyong Huang", "Tat-Seng Chua"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "679bfcac736544e9a241c579adfda0cacbc40838452ea56ad631f8086868415a", "submitted_at": "2025-04-04T14:25:17Z", "updated_at": "2025-04-04T14:25:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.07981v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.07981", "identifiers": {"arxiv": "2504.07981"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.07981"}], "name": "ScreenSpot-Pro: GUI Grounding for Professional High-Resolution Computer Use", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-04-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates GUI target localization in high-resolution professional software and studies cascaded visual search to narrow the target region.", "topics": ["grounding", "evaluation"], "updated_at": "2025-04-04", "url": "https://arxiv.org/abs/2504.07981"}
{"aliases": [], "authors": ["Xing Han Lù", "Amirhossein Kazemnejad", "Nicholas Meade", "Arkil Patel", "Dongchan Shin", "Alejandra Zambrano", "Karolina Stańczak", "Peter Shaw", "Christopher J. Pal", "Siva Reddy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53cba694577319ac1992c853be79735d654d9d32c12edcc543e6738ce010050d", "submitted_at": "2025-04-11T19:49:22Z", "updated_at": "2025-10-06T15:46:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.08942v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.08942", "identifiers": {"arxiv": "2504.08942"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.08942"}], "name": "AgentRewardBench: Evaluating Automatic Evaluations of Web Agent Trajectories", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "AgentRewardBench evaluates automated web-agent judges against expert reviews of 1,302 trajectories from five benchmarks. The annotations cover completion, side effects, and repetition, enabling comparison of LLM judges and rule-based task evaluators.", "topics": ["evaluation", "web-navigation"], "updated_at": "2025-10-06", "url": "https://arxiv.org/abs/2504.08942"}
{"aliases": [], "authors": ["Iason Chaimalas", "Arnas Vyšniauskas", "Gabriel Brostow"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf23700bf7626733b0016b773852455c503617adc81025fc9e097879b5534cca", "submitted_at": "2025-04-12T22:02:29Z", "updated_at": "2025-04-12T22:02:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.09352v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "id": "arxiv:2504.09352", "identifiers": {"arxiv": "2504.09352"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.09352"}], "name": "Explorer: Robust Collection of Interactable GUI Elements", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Explorer collects interactable GUI elements from a live target application and records user sessions for application-specific model training. It links repeated interface states into a navigation map that supports path planning and voice-directed interaction.", "topics": ["data-generation", "recording", "grounding", "planning"], "updated_at": "2025-04-12", "url": "https://arxiv.org/abs/2504.09352"}
{"aliases": [], "authors": ["Yuxuan Lu", "Ting-Yao Hsu", "Hansu Gu", "Limeng Cui", "Yaochen Xie", "William Headden", "Bingsheng Yao", "Akash Veeragouni", "Jiapeng Liu", "Sreyashi Nag", "Jessie Wang", "Dakuo Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "743ccb4cd6202f29784fe6df4cb0b46f6ca9e16c429cbd4cb48cbad99a8fd4d7", "submitted_at": "2025-04-13T21:10:56Z", "updated_at": "2026-03-10T20:51:15Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.09723v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.09723", "identifiers": {"arxiv": "2504.09723"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.09723"}], "name": "AgentA/B: Automated and Scalable Web A/BTesting with Interactive LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-10", "url": "https://arxiv.org/abs/2504.09723"}
{"aliases": [], "authors": ["Junlei Zhang", "Zichen Ding", "Chang Ma", "Zijie Chen", "Qiushi Sun", "Zhenzhong Lan", "Junxian He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "171517999164a3581f71cf363ec6eed357b28ca86908cb97e8498c6958762a5f", "submitted_at": "2025-04-14T11:35:02Z", "updated_at": "2025-04-15T17:13:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.10127v2"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.10127", "identifiers": {"arxiv": "2504.10127"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.10127"}], "name": "Breaking the Data Barrier -- Building GUI Agents Through Task Generalization", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies whether mid-training on GUI perception, multimodal reasoning, and text reasoning transfers to GUI-agent planning. Experiments compare 11 task sources and curated mixtures using WebArena and AndroidWorld to investigate training-data choices when GUI trajectories are scarce.", "topics": ["data-generation", "planning", "evaluation"], "updated_at": "2025-04-15", "url": "https://arxiv.org/abs/2504.10127"}
{"aliases": [], "authors": ["Xiaopeng Li", "Pengyue Jia", "Derong Xu", "Yi Wen", "Yingyi Zhang", "Wenlin Zhang", "Wanyu Wang", "Yichao Wang", "Zhaocheng Du", "Xiangyang Li", "Yong Liu", "Huifeng Guo", "Ruiming Tang", "Xiangyu Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ef2f51751f2a79bafec208f023d373655ccd5def2c6386bbc6c1507150ea77cc", "submitted_at": "2025-04-14T11:57:52Z", "updated_at": "2025-04-14T11:57:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.10147v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2504.10147", "identifiers": {"arxiv": "2504.10147"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.10147"}], "name": "A Survey of Personalization: From RAG to Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-14", "url": "https://arxiv.org/abs/2504.10147"}
{"aliases": [], "authors": ["Suyu Ye", "Haojun Shi", "Darren Shih", "Hyokun Yun", "Tanya Roosta", "Tianmin Shu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "268adf347e1f224f434e6dba838873c99a4050686cae1495643d9638d840d809", "submitted_at": "2025-04-14T17:36:46Z", "updated_at": "2025-12-01T15:22:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.10445v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2504.10445", "identifiers": {"arxiv": "2504.10445"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.10445"}], "name": "RealWebAssist: A Benchmark for Long-Horizon Web Assistance with Real-World Users", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "RealWebAssist evaluates web assistance through sequences of instructions collected from real users across multiple websites. It tests intent interpretation, user-specific routines, long-term context, and visual grounding when instructions are ambiguous or evolve during interaction.", "topics": ["evaluation", "long-horizon", "human-in-the-loop", "grounding"], "updated_at": "2025-12-01", "url": "https://arxiv.org/abs/2504.10445"}
{"aliases": [], "authors": ["Run Luo", "Lu Wang", "Wanwei He", "Longze Chen", "Jiaming Li", "Xiaobo Xia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b875617b795e0f8d7d6348e22ec85ccc9ab4dd931e66850182348927b1e5209f", "submitted_at": "2025-04-14T17:45:54Z", "updated_at": "2025-10-01T04:55:20Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.10458v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.10458", "identifiers": {"arxiv": "2504.10458"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.10458"}], "name": "GUI-R1 : A Generalist R1-Style Vision-Language Action Model For GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Applies reinforcement fine-tuning with unified action rules and curated cross-platform examples to vision-language GUI agents.", "topics": ["reinforcement-learning", "grounding"], "updated_at": "2025-10-01", "url": "https://arxiv.org/abs/2504.10458"}
{"aliases": [], "authors": ["Xinyi Liu", "Xiaoyi Zhang", "Ziyun Zhang", "Yan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "363e3c3e882fa93a19df74bbf77aa98525cf4639c8f62100c3ddc3e0a438997f", "submitted_at": "2025-04-15T14:56:21Z", "updated_at": "2025-07-30T13:57:49Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.11257v4"}, "publication": {"authors": ["Xinyi Liu", "Xiaoyi Zhang", "Ziyun Zhang", "Yan Lu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.809", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.809/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "15668-15684", "pdf_url": "https://aclanthology.org/2025.findings-acl.809.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "UI-E2I-Synth: Advancing GUI Grounding with Large-Scale Instruction Synthesis", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.809/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:59Z", "role": "primary", "sha256": "f7bcea2d3656ebc3ac305d65271000e411b735eb10556d23ee0d6157593a109f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.809/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.11257", "identifiers": {"arxiv": "2504.11257"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.11257"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.809/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.809.pdf"}], "name": "UI-E2I-Synth: Advancing GUI Grounding with Large-Scale Instruction Synthesis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-E2I-Synth uses GPT-4o to synthesize GUI grounding instructions with varied complexity, addressing small elements, imbalanced element types, and implicit requests. The accompanying UI-I2E-Bench evaluates instruction-to-element localization across these annotation challenges.", "topics": ["data-generation", "grounding", "evaluation"], "updated_at": "2025-07-30", "url": "https://arxiv.org/abs/2504.11257"}
{"aliases": [], "authors": ["Chaoran Chen", "Zhiping Zhang", "Bingcan Guo", "Shang Ma", "Ibrahim Khalilov", "Simret A Gebreegziabher", "Yanfang Ye", "Ziang Xiao", "Yaxing Yao", "Tianshi Li", "Toby Jia-Jun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "48acf3279b854d45ec076e630e5a84601879fc9ff894c5d368adb59d03bd3de6", "submitted_at": "2025-04-15T15:21:09Z", "updated_at": "2025-04-15T15:21:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.11281v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.11281", "identifiers": {"arxiv": "2504.11281"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.11281"}], "name": "The Obvious Invisible Threat: LLM-Powered GUI Agents' Vulnerability to Fine-Print Injections", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Characterizes six forms of GUI-content injection and studies their effects on GUI agents and human users using adversarial webpages. The analysis examines privacy failures and the limits of human oversight.", "topics": ["safety-security", "human-in-the-loop", "evaluation"], "updated_at": "2025-04-15", "url": "https://arxiv.org/abs/2504.11281"}
{"aliases": [], "authors": ["Yangyang Zhuang", "Wenjia Jiang", "Jiayu Zhang", "Ze Yang", "Joey Tianyi Zhou", "Chi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c07281bb36ab8b3cdb71bcbee02c8656f104fbff1c5ef747cd5f84108d5386c1", "submitted_at": "2025-04-15T15:44:21Z", "updated_at": "2025-08-15T08:59:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.11301v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.11301", "identifiers": {"arxiv": "2504.11301"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.11301"}], "name": "Learning to Be A Doctor: Searching for Effective Medical Agent Architectures", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-15", "url": "https://arxiv.org/abs/2504.11301"}
{"aliases": [], "authors": ["Divyansh Garg", "Shaun VanWeelden", "Diego Caples", "Andis Draguns", "Nikil Ravi", "Pranav Putta", "Naman Garg", "Tomas Abraham", "Michael Lara", "Federico Lopez", "James Liu", "Atharva Gundawar", "Prannay Hebbar", "Youngchul Joo", "Jindong Gu", "Charles London", "Christian Schroeder de Witt", "Sumeet Motwani"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8fe0d8b34d167311d1c37934fdb06e81939652ea579b2578a40b6a709fbfb91", "submitted_at": "2025-04-15T18:22:55Z", "updated_at": "2025-04-17T16:28:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.11543v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.11543", "identifiers": {"arxiv": "2504.11543"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.11543"}], "name": "REAL: Benchmarking Autonomous Agents on Deterministic Simulations of Real Websites", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-17", "url": "https://arxiv.org/abs/2504.11543"}
{"aliases": [], "authors": ["Biniam Fisseha Demissie", "Yan Naing Tun", "Lwin Khin Shar", "Mariano Ceccato"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a25cb3ca680732a08cf2c461a65caa0c6a70404224bdfd919ba5bec7760ba72", "submitted_at": "2025-04-16T00:19:31Z", "updated_at": "2025-04-16T00:19:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.11675v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2504.11675", "identifiers": {"arxiv": "2504.11675"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.11675"}], "name": "VLM-Fuzz: Vision Language Model Assisted Recursive Depth-first Search Exploration for Effective UI Testing of Android Apps", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "VLM-Fuzz combines depth-first UI exploration with Android manifest analysis, runtime widget hierarchies, and on-demand visual reasoning to generate app test inputs. It evaluates coverage on 59 benchmark apps and uses additional Google Play apps to study crash discovery.", "topics": ["automation", "evaluation", "mobile-navigation"], "updated_at": "2025-04-16", "url": "https://arxiv.org/abs/2504.11675"}
{"aliases": [], "authors": ["Tianneng Shi", "Jingxuan He", "Zhun Wang", "Hongwei Li", "Linyu Wu", "Wenbo Guo", "Dawn Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2613f1b3e60cf7fd81d320bcb9d47841562992e0e2ea3b5b8d969656317d6d74", "submitted_at": "2025-04-16T01:58:40Z", "updated_at": "2026-05-14T02:36:38Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.11703v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2504.11703", "identifiers": {"arxiv": "2504.11703"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.11703"}], "name": "Progent: Securing AI Agents with Privilege Control", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-14", "url": "https://arxiv.org/abs/2504.11703"}
{"aliases": [], "authors": ["Zhisong Zhang", "Tianqing Fang", "Kaixin Ma", "Wenhao Yu", "Hongming Zhang", "Haitao Mi", "Dong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76f84d5375ab7321512a1f4f8724c011cf180bf5a7a0d532ff3218cc9d7a6e75", "submitted_at": "2025-04-16T05:41:20Z", "updated_at": "2026-01-15T02:00:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.11788v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.11788", "identifiers": {"arxiv": "2504.11788"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.11788"}], "name": "WebRollback: Enhancing Web Agents with Explicit Rollback Mechanisms", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebRollback gives a web agent an explicit action for returning to an earlier state in its navigation trajectory. It studies this recovery mechanism on two live web-navigation benchmarks under both zero-shot and fine-tuned settings.", "topics": ["web-navigation", "planning", "robustness"], "updated_at": "2026-01-15", "url": "https://arxiv.org/abs/2504.11788"}
{"aliases": [], "authors": ["Tyler Ga Wei Lum", "Olivia Y. Lee", "C. Karen Liu", "Jeannette Bohg"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5522406f7f44bccf72e7c079b5287ed0bff2ab208104a4129be03e03240459a0", "submitted_at": "2025-04-17T03:15:20Z", "updated_at": "2025-08-16T04:32:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.12609v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.12609", "identifiers": {"arxiv": "2504.12609"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.12609"}], "name": "Crossing the Human-Robot Embodiment Gap with Sim-to-Real RL using One Human Demonstration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-16", "url": "https://arxiv.org/abs/2504.12609"}
{"aliases": [], "authors": ["Bofei Zhang", "Zirui Shang", "Zhi Gao", "Wang Zhang", "Rui Xie", "Xiaojian Ma", "Tao Yuan", "Xinxiao Wu", "Song-Chun Zhu", "Qing Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "346fc4db7f8fbf2dc66c32e36f5cc5ace22228d60bdd4c61b6b7ac65487a4a8c", "submitted_at": "2025-04-17T06:15:56Z", "updated_at": "2025-12-04T15:01:09Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.12679v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.12679", "identifiers": {"arxiv": "2504.12679"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.12679"}], "name": "TongUI: Internet-Scale Trajectories from Multimodal Web Tutorials for Generalized GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "TongUI converts multimodal online tutorials, including videos and articles, into GUI-agent trajectories. Its GUI-Net dataset contains 143,000 trajectories spanning five operating systems and more than 200 applications, and supports fine-tuning for grounding and navigation.", "topics": ["data-generation", "grounding", "video"], "updated_at": "2025-12-04", "url": "https://arxiv.org/abs/2504.12679"}
{"aliases": [], "authors": ["Guangyi Liu", "Pengxiang Zhao", "Liang Liu", "Zhiming Chen", "Yuxiang Chai", "Shuai Ren", "Hao Wang", "Shibo He", "Wenchao Meng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bbe6c4969a9557546752235809d25912bff96a97001fade30102d8d2c751054", "submitted_at": "2025-04-18T17:13:34Z", "updated_at": "2025-04-18T17:13:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.13805v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.13805", "identifiers": {"arxiv": "2504.13805"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.13805"}], "name": "LearnAct: Few-Shot Mobile GUI Agent with a Unified Demonstration Benchmark", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "LearnAct parses human demonstrations into reusable knowledge, retrieves relevant examples, and uses them during mobile GUI task execution. The accompanying LearnGUI benchmark includes 2,252 offline tasks and 101 online tasks for studying adaptation from demonstrations.", "topics": ["mobile-navigation", "memory", "human-in-the-loop", "evaluation"], "updated_at": "2025-04-18", "url": "https://arxiv.org/abs/2504.13805"}
{"aliases": [], "authors": ["Fei Tang", "Haolei Xu", "Hang Zhang", "Siqi Chen", "Xingyu Wu", "Yongliang Shen", "Wenqi Zhang", "Guiyang Hou", "Zeqi Tan", "Yuchen Yan", "Kaitao Song", "Jian Shao", "Weiming Lu", "Jun Xiao", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "216d0605205cfd7f17ffd7d7f4c9752577ca427f370df42faf4a2250b80bcc39", "submitted_at": "2025-03-27T17:58:31Z", "updated_at": "2025-06-04T17:29:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.13865v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.13865", "identifiers": {"arxiv": "2504.13865"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.13865"}], "name": "A Survey on (M)LLM-Based GUI Agents", "notes": "", "published_at": "2025-03-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes LLM-based GUI agents through four components (perception, exploration, planning, and interaction) and the methods used to evaluate them.", "topics": ["survey"], "updated_at": "2025-06-04", "url": "https://arxiv.org/abs/2504.13865"}
{"aliases": [], "authors": ["Dezhao Luo", "Bohan Tang", "Kang Li", "Georgios Papoudakis", "Jifei Song", "Shaogang Gong", "Jianye Hao", "Jun Wang", "Kun Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15e00237ef710c37c0c9f7768d699f57e9d5dcc190e8de7e20e9ec3b21b82750", "submitted_at": "2025-04-15T14:03:10Z", "updated_at": "2025-05-20T14:11:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.13936v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "id": "arxiv:2504.13936", "identifiers": {"arxiv": "2504.13936"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.13936"}], "name": "ViMo: A Generative Visual GUI World Model for App Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "ViMo predicts the next mobile-app screenshot after a candidate action by separately modeling graphics and text. A symbolic placeholder representation preserves layout while a second predictor supplies readable text, enabling visual lookahead for GUI-agent planning.", "topics": ["world-models", "planning", "mobile-navigation"], "updated_at": "2025-05-20", "url": "https://arxiv.org/abs/2504.13936"}
{"aliases": [], "authors": ["Leo Boisvert", "Mihir Bansal", "Chandra Kiran Reddy Evuru", "Gabriel Huang", "Abhay Puri", "Avinandan Bose", "Maryam Fazel", "Quentin Cappart", "Jason Stanley", "Alexandre Lacoste", "Alexandre Drouin", "Krishnamurthy Dvijotham"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5bdf5fefa1ee31126d99de08236991df559896212f60e0b3058fcd8cc8a134b", "submitted_at": "2025-04-18T20:36:10Z", "updated_at": "2025-10-07T13:10:30Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.14064v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2504.14064", "identifiers": {"arxiv": "2504.14064"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.14064"}], "name": "DoomArena: A framework for Testing AI Agents Against Evolving Security Threats", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents a modular security testing framework that plugs into BrowserGym and tool-use benchmarks to evaluate agents under configurable threat models.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-10-07", "url": "https://arxiv.org/abs/2504.14064"}
{"aliases": [], "authors": ["Yong-En Tian", "Yu-Chien Tang", "Kuang-Da Wang", "An-Zi Yen", "Wen-Chih Peng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b35c325b5dc4d3261838dea86d40a6a18a98364c8fd0197279ee9629ddd9bb9a", "submitted_at": "2025-04-19T09:06:13Z", "updated_at": "2025-04-19T09:06:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.14233v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.14233", "identifiers": {"arxiv": "2504.14233"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.14233"}], "name": "Template-Based Financial Report Generation in Agentic and Decomposed Information Retrieval", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-19", "url": "https://arxiv.org/abs/2504.14233"}
{"aliases": [], "authors": ["Yuhang Liu", "Pengxiang Li", "Congkai Xie", "Xavier Hu", "Xiaotian Han", "Shengyu Zhang", "Hongxia Yang", "Fei Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ce03185cc73c2723f444bd9729c67da9aab7d96b5ef099fb78e47f58790bc18", "submitted_at": "2025-04-19T09:25:55Z", "updated_at": "2025-04-19T09:25:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.14239v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.14239", "identifiers": {"arxiv": "2504.14239"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.14239"}], "name": "InfiGUI-R1: Advancing Multimodal GUI Agents from Reactive Actors to Deliberative Reasoners", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-04-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses a staged reasoning-focused training framework to move GUI agents from reactive actions toward deliberative planning and recovery.", "topics": ["planning", "reinforcement-learning"], "updated_at": "2025-04-19", "url": "https://arxiv.org/abs/2504.14239"}
{"aliases": [], "authors": ["Le Wang", "Zonghao Ying", "Tianyuan Zhang", "Siyuan Liang", "Shengshan Hu", "Mingchuan Zhang", "Aishan Liu", "Xianglong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "481077817953a688409b0422d94976011077e501479fd127d48fa67b02a909ac", "submitted_at": "2025-04-19T16:28:03Z", "updated_at": "2025-07-27T10:12:52Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.14348v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.14348", "identifiers": {"arxiv": "2504.14348"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.14348"}], "name": "Manipulating Multimodal Agents via Cross-Modal Prompt Injection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-27", "url": "https://arxiv.org/abs/2504.14348"}
{"aliases": [], "authors": ["Chaoyun Zhang", "He Huang", "Chiming Ni", "Jian Mu", "Si Qin", "Shilin He", "Lu Wang", "Fangkai Yang", "Pu Zhao", "Chao Du", "Liqun Li", "Yu Kang", "Zhao Jiang", "Suzhen Zheng", "Rujia Wang", "Jiaxu Qian", "Minghua Ma", "Jian-Guang Lou", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc47a80b48a6a5f126e36fa14639d56d8c8025a453fbcee2cbdb73eb67945561", "submitted_at": "2025-04-20T13:04:43Z", "updated_at": "2025-04-25T05:14:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.14603v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:07Z", "role": "primary", "sha256": "0ff3084c62a2ae0fbda31be5d409fb8ca3a6887da897a238beec93a8d7be7825", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2504.14603v2"}], "id": "arxiv:2504.14603", "identifiers": {"arxiv": "2504.14603"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:07Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.14603"}], "name": "UFO2: The Desktop AgentOS", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-04-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces UFO2, a Windows desktop agent architecture combining a coordinating HostAgent, application-specialized agents, and a shared GUI/API action layer. It integrates UI Automation, visual parsing, speculative planning, and an isolated picture-in-picture desktop.", "topics": ["automation", "cross-app", "tool-integration"], "updated_at": "2025-04-25", "url": "https://arxiv.org/abs/2504.14603"}
{"aliases": [], "authors": ["Zhiyuan Hu", "Shiyun Xiong", "Yifan Zhang", "See-Kiong Ng", "Anh Tuan Luu", "Bo An", "Shuicheng Yan", "Bryan Hooi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1353bb4651f449763ac58803b4858edf6e1d5b0ad3ab3e4c333883a1ebbaebf4", "submitted_at": "2025-04-22T17:52:42Z", "updated_at": "2025-04-22T17:52:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.16073v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:10Z", "role": "primary", "sha256": "b1cf161fc20efc51363d9007952ce5f917963d4ab2ca32d39099229060b29528", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2504.16073v1"}], "id": "arxiv:2504.16073", "identifiers": {"arxiv": "2504.16073"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.16073"}], "name": "Guiding VLM Agents with Process Rewards at Inference Time for GUI Navigation", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-04-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Guides visual GUI agents at inference time with a process-reward model that evaluates intermediate actions. It combines stepwise guidance with trajectory reflection and retry mechanisms in static and interactive navigation tasks.", "topics": ["reward-models", "self-reflection", "grounding"], "updated_at": "2025-04-22", "url": "https://arxiv.org/abs/2504.16073"}
{"aliases": [], "authors": ["Qi Yang", "Weichen Bi", "Haiyang Shen", "Yaoqi Guo", "Yun Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1ec3311094166baac4de66ad0f083b66585c933f19cb66fc580bf128ad809ca", "submitted_at": "2025-04-23T05:01:25Z", "updated_at": "2025-04-27T06:52:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.16419v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:13Z", "role": "primary", "sha256": "496c0419e2c8c752fc2c88d015ad2cf13c182aa4501a8305ff65ad4178113843", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2504.16419v2"}], "id": "arxiv:2504.16419", "identifiers": {"arxiv": "2504.16419"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.16419"}], "name": "PixelWeb: The First Web GUI Dataset with Pixel-Wise Labels", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-04-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PixelWeb, a webpage GUI dataset with element images, contours, and pixel masks. Its annotation pipeline combines bitmap channel information with DOM-based visibility and stacking analysis to address occlusion and overlapping elements.", "topics": ["data-generation", "grounding", "screen-understanding"], "updated_at": "2025-04-27", "url": "https://arxiv.org/abs/2504.16419"}
{"aliases": [], "authors": ["Yasmine Omri", "Parth Shroff", "Thierry Tambe"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c5a5b902b29e3b1a74ef79eda1ec9198aa87b757fb53fd9df41c451de0904b2", "submitted_at": "2025-04-24T19:11:10Z", "updated_at": "2025-04-24T19:11:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.17892v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.17892", "identifiers": {"arxiv": "2504.17892"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.17892"}], "name": "Token Sequence Compression for Efficient Multimodal Computing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-24", "url": "https://arxiv.org/abs/2504.17892"}
{"aliases": [], "authors": ["Chaoran Chen", "Zhiping Zhang", "Ibrahim Khalilov", "Bingcan Guo", "Simret A Gebreegziabher", "Yanfang Ye", "Ziang Xiao", "Yaxing Yao", "Tianshi Li", "Toby Jia-Jun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2a55e347bc10e5571464924d9301ff2c32d5e9f1f990ca91725ec8c9f6da594", "submitted_at": "2025-04-24T20:51:20Z", "updated_at": "2025-06-05T12:40:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.17934v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.17934", "identifiers": {"arxiv": "2504.17934"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.17934"}], "name": "Toward a Human-Centered Evaluation Framework for Trustworthy LLM-Powered GUI Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. This is a position paper proposing an evaluation framework; the abstract does not report a new benchmark or empirical validation of the proposed framework.", "published_at": "2025-04-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents a position paper on privacy and security evaluation for GUI agents, identifying risk categories and challenges in involving human evaluators. It proposes risk-aware assessment and in-context consent as design directions.", "topics": ["evaluation", "safety-security", "human-in-the-loop"], "updated_at": "2025-06-05", "url": "https://arxiv.org/abs/2504.17934"}
{"aliases": [], "authors": ["Ivan Evtimov", "Arman Zharmagambetov", "Aaron Grattafiori", "Chuan Guo", "Kamalika Chaudhuri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a104d9076797555795766749f6b4eeec9428d0efb9340f7e55cffecd5d727b0", "submitted_at": "2025-04-22T17:51:03Z", "updated_at": "2025-05-16T22:42:29Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.18575v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.18575", "identifiers": {"arxiv": "2504.18575"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.18575"}], "name": "WASP: Benchmarking Web Agent Security Against Prompt Injection Attacks", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-04-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WASP, an end-to-end benchmark for prompt-injection security in web agents. Its evaluation separates partial attack progress from completion of the attacker’s full objective.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2025-05-16", "url": "https://arxiv.org/abs/2504.18575"}
{"aliases": [], "authors": ["Shi Jie Yu", "Sehyun Choi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f9e5e381734feebcf370554e06f075ed9eed5180d914a781104b65f74e229ab7", "submitted_at": "2025-04-23T05:11:21Z", "updated_at": "2025-04-23T05:11:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.18580v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.18580", "identifiers": {"arxiv": "2504.18580"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.18580"}], "name": "Parameter-Efficient Checkpoint Merging via Metrics-Weighted Averaging", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-23", "url": "https://arxiv.org/abs/2504.18580"}
{"aliases": [], "authors": ["Guangyi Liu", "Pengxiang Zhao", "Yaozhen Liang", "Liang Liu", "Yaxuan Guo", "Han Xiao", "Weifeng Lin", "Yuxiang Chai", "Yue Han", "Shuai Ren", "Hao Wang", "Xiaoyu Liang", "WenHao Wang", "Tianze Wu", "Zhengxi Lu", "Siheng Chen", "LiLinghao", "Hao Wang", "Guanjing Xiong", "Yong Liu", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9103197fa1274fac79ae9e4472f5fbdc9a210548afb024f1d4259c6bb0131bf2", "submitted_at": "2025-04-28T14:39:25Z", "updated_at": "2025-11-17T15:51:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.19838v3"}}, "environments": ["android", "ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.19838", "identifiers": {"arxiv": "2504.19838"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.19838"}], "name": "LLM-Powered GUI Agents in Phone Automation: Surveying Progress and Prospects", "notes": "", "published_at": "2025-04-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Reviews LLM-driven phone GUI agents with a taxonomy of agent frameworks, prompt- and training-based modeling, datasets, and benchmarks.", "topics": ["survey"], "updated_at": "2025-11-17", "url": "https://arxiv.org/abs/2504.19838"}
{"aliases": [], "authors": ["Jiahao Li", "Kaer Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5040073e7c278734df219d71d94e236468468621f5971ef3ccbc97b94e1792b9", "submitted_at": "2025-04-29T06:55:15Z", "updated_at": "2025-05-13T01:59:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.20464v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.20464", "identifiers": {"arxiv": "2504.20464"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.20464"}], "name": "A Survey on GUI Agents with Foundation Models Enhanced by Reinforcement Learning", "notes": "", "published_at": "2025-04-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Formalizes GUI tasks as Markov decision processes and traces agent training from prompting through supervised fine-tuning to reinforcement learning.", "topics": ["survey", "reinforcement-learning"], "updated_at": "2025-05-13", "url": "https://arxiv.org/abs/2504.20464"}
{"aliases": [], "authors": ["Zikui Cai", "Shayan Shabihi", "Bang An", "Zora Che", "Brian R. Bartoldson", "Bhavya Kailkhura", "Tom Goldstein", "Furong Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd8cdbba7db9431e44603bc2050eb6e502dea4ac3b166812e11550fadd96abc4", "submitted_at": "2025-04-29T17:36:05Z", "updated_at": "2025-06-13T22:43:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.20965v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2504.20965", "identifiers": {"arxiv": "2504.20965"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2504.20965"}], "name": "AegisLLM: Scaling Agentic Systems for Self-Reflective Defense in LLM Security", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-13", "url": "https://arxiv.org/abs/2504.20965"}
{"aliases": [], "authors": ["Tianqing Fang", "Hongming Zhang", "Zhisong Zhang", "Kaixin Ma", "Wenhao Yu", "Haitao Mi", "Dong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86ca58c92ffc2cbbcb0297d8c7618dfab966e36a07b141c068998b0f2934d11e", "submitted_at": "2025-04-23T02:54:31Z", "updated_at": "2025-08-21T05:55:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.21024v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2504.21024", "identifiers": {"arxiv": "2504.21024"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.21024"}], "name": "WebEvolver: Enhancing Web Agent Self-Improvement with Coevolving World Model", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-04-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebEvolver trains a co-evolving world model to predict web observations and generate additional agent-training trajectories. The same model supports look-ahead action selection at inference, with evaluation on live web-task benchmarks.", "topics": ["world-models", "data-generation", "web-navigation", "planning"], "updated_at": "2025-08-21", "url": "https://arxiv.org/abs/2504.21024"}
{"aliases": [], "authors": ["Xiaoxi Li", "Jiajie Jin", "Guanting Dong", "Hongjin Qian", "Yongkang Wu", "Ji-Rong Wen", "Yutao Zhu", "Zhicheng Dou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "351b6d7fbe4e217a5211884832c6ad07d76a0463e6a8c996815999ff3a5f1921", "submitted_at": "2025-04-30T16:25:25Z", "updated_at": "2025-10-13T12:40:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.21776v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T13:00:06.851520+00:00", "role": "primary", "sha256": "d06baa20c7775347cbb4492473418d11cbe56941ab90baa16381353ab6c39e07", "supports": ["summary", "scope", "notes"], "url": "https://arxiv.org/html/2504.21776v2"}], "id": "arxiv:2504.21776", "identifiers": {"arxiv": "2504.21776"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.21776"}], "name": "WebThinker: Empowering Large Reasoning Models with Deep Research Capability", "notes": "Adjacent web-research method. The retained abstract and full-text Section 3 and Appendix A describe search APIs and Crawl4AI URL retrieval; no visual GUI execution is inferred. Results were not independently reproduced.", "published_at": "2025-04-30", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Combines iterative web search and link-following with interleaved reasoning and report drafting for deep research. Its reported web explorer retrieves page content through a crawler rather than controlling a visual browser interface.", "topics": ["web-navigation", "planning", "tool-integration"], "updated_at": "2025-10-13", "url": "https://arxiv.org/abs/2504.21776"}
{"aliases": [], "authors": ["Zheng Zhang", "Jinyi Li", "Yihuai Lan", "Xiang Wang", "Hao Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "678919c145c11002704b37fd7b2d68c3ea51d29e0933799ae2719bde6e16634c", "submitted_at": "2025-04-24T14:15:13Z", "updated_at": "2025-04-24T14:15:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.00019v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.00019", "identifiers": {"arxiv": "2505.00019"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.00019"}], "name": "An Empirical Study on Prompt Compression for Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-04-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-04-24", "url": "https://arxiv.org/abs/2505.00019"}
{"aliases": [], "authors": ["Jing Huang", "Zhixiong Zeng", "Wenkang Han", "Yufeng Zhong", "Liming Zheng", "Shuai Fu", "Jingyuan Chen", "Lin Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e582c92521ec41a0a3daa0ca72b8c00250621289352b28400ee6d311c4a106eb", "submitted_at": "2025-05-01T09:27:13Z", "updated_at": "2025-05-01T09:27:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.00416v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.00416", "identifiers": {"arxiv": "2505.00416"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.00416"}], "name": "ScaleTrack: Scaling and back-tracking Automated GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "ScaleTrack combines a unified collection of GUI grounding samples with a planning objective that predicts the next action while reconstructing prior actions. This backward prediction links the current screenshot to interaction history during agent training.", "topics": ["grounding", "planning", "data-generation", "interaction-history"], "updated_at": "2025-05-01", "url": "https://arxiv.org/abs/2505.00416"}
{"aliases": [], "authors": ["Tiange Luo", "Lajanugen Logeswaran", "Justin Johnson", "Honglak Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d2dc8ca397adce1ecf6cf53afa82a8d4f1825bf00277aeb117966f965abfe48", "submitted_at": "2025-05-01T17:45:59Z", "updated_at": "2025-07-14T17:45:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.00684v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.00684", "identifiers": {"arxiv": "2505.00684"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.00684"}], "name": "Visual Test-time Scaling for GUI Agent Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "RegionFocus adds visual test-time scaling by zooming into selected screenshot regions and maintaining an image-based map of landmarks and actions. It evaluates this grounding procedure with UI-TARS and Qwen2.5-VL on ScreenSpot-Pro and WebVoyager.", "topics": ["grounding", "screen-understanding", "web-navigation"], "updated_at": "2025-07-14", "url": "https://arxiv.org/abs/2505.00684"}
{"aliases": [], "authors": ["Yanbang Li", "Ziyang Gong", "Haoyang Li", "Xiaoqi Huang", "Haolan Kang", "Guangping Bai", "Xianzheng Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8682f83fa75a960b8c668b46d5eaa4b64746f1cac6bb9b30ddabc81e5619bcf8", "submitted_at": "2025-05-01T17:55:05Z", "updated_at": "2025-07-27T09:58:24Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.00693v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.00693", "identifiers": {"arxiv": "2505.00693"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.00693"}], "name": "Robotic Visual Instruction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-27", "url": "https://arxiv.org/abs/2505.00693"}
{"aliases": [], "authors": ["Gabriel Sarch", "Balasaravanan Thoravi Kumaravel", "Sahithya Ravi", "Vibhav Vineet", "Andrew D. Wilson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f9b897b1d7f1b7dc2cf16fbf8e0b3b6a948c3115ed53d0c19a991c76810a87e", "submitted_at": "2025-05-02T20:43:11Z", "updated_at": "2025-05-02T20:43:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.01578v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.01578", "identifiers": {"arxiv": "2505.01578"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.01578"}], "name": "Grounding Task Assistance with Multimodal Cues from a Single Demonstration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-02", "url": "https://arxiv.org/abs/2505.01578"}
{"aliases": [], "authors": ["Runyi Yu", "Yinhuai Wang", "Qihan Zhao", "Hok Wai Tsui", "Jingbo Wang", "Ping Tan", "Qifeng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2145e32afe5b3950a09249982fa2637819d8fc4374ed152c0c12de8bb31f8bca", "submitted_at": "2025-05-04T13:00:29Z", "updated_at": "2025-05-04T13:00:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.02094v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.02094", "identifiers": {"arxiv": "2505.02094"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.02094"}], "name": "SkillMimic-V2: Learning Robust and Generalizable Interaction Skills from Sparse and Noisy Demonstrations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-04", "url": "https://arxiv.org/abs/2505.02094"}
{"aliases": [], "authors": ["Mariya Davydova", "Daniel Jeffries", "Patrick Barker", "Arturo Márquez Flores", "Sinéad Ryan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8bb4c6f8080ed1ee91e1089233cae6f231ab3256b38ac7d6403ebea3ae56fe0", "submitted_at": "2025-05-06T14:29:47Z", "updated_at": "2025-05-06T14:29:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.03570v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:17Z", "role": "primary", "sha256": "39de8f4a9aee74cc0eae64e24efc08ed070714eaaea4a30dbc20cb08397eb9bd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2505.03570v1"}], "id": "arxiv:2505.03570", "identifiers": {"arxiv": "2505.03570"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.03570"}], "name": "OSUniverse: Benchmark for Multimodal GUI-navigation AI Agents", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-05-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces OSUniverse, a desktop GUI benchmark spanning precise clicks through multistep and multiapplication tasks. It provides both manual scoring and an automated validation mechanism.", "topics": ["evaluation", "cross-app", "long-horizon"], "updated_at": "2025-05-06", "url": "https://arxiv.org/abs/2505.03570"}
{"aliases": [], "authors": ["Sahana Chennabasappa", "Cyrus Nikolaidis", "Daniel Song", "David Molnar", "Stephanie Ding", "Shengye Wan", "Spencer Whitman", "Lauren Deason", "Nicholas Doucette", "Abraham Montilla", "Alekhya Gampa", "Beto de Paola", "Dominik Gabi", "James Crnkovich", "Jean-Christophe Testud", "Kat He", "Rashnil Chaturvedi", "Wu Zhou", "Joshua Saxe"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e4a9b47f999eb820e574ed8cb1023379d57f2f891a16059d9820b8c6942ec3f", "submitted_at": "2025-05-06T14:34:21Z", "updated_at": "2025-05-06T14:34:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.03574v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.03574", "identifiers": {"arxiv": "2505.03574"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.03574"}], "name": "LlamaFirewall: An open source guardrail system for building secure AI agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-06", "url": "https://arxiv.org/abs/2505.03574"}
{"aliases": [], "authors": ["Jaehyun Jeon", "Min Soo Kim", "Jang Han Yoon", "Sumin Shim", "Yejin Choi", "Hanbin Kim", "Dae Hyun Kim", "Youngjae Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f65cd2856c554584baab9b1cbf9c5250b5c01d4f31980d69aefab727427ded2", "submitted_at": "2025-05-08T08:00:32Z", "updated_at": "2026-06-04T06:58:13Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2505.05026v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:21Z", "role": "primary", "sha256": "1220bb1498a7f9c68cff0fbbd8eee209dfed138e771dfd6eb946fc0639c49cd1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2505.05026v5"}], "id": "arxiv:2505.05026", "identifiers": {"arxiv": "2505.05026"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.05026"}], "name": "Do MLLMs Capture How Interfaces Guide User Behavior? A Benchmark for Multimodal UI/UX Design Understanding", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced. This benchmark measures UI/UX design understanding from static image pairs; it does not measure autonomous GUI action execution.", "published_at": "2025-05-08", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces WiserUI-Bench, pairing real UI variants from industry A/B tests with observed behavioral winners and expert explanations. It evaluates models’ ability to predict the more effective interface and explain its effect on user behavior.", "topics": ["screen-understanding", "ui-design", "evaluation"], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2505.05026"}
{"aliases": [], "authors": ["Biao Yi", "Xavier Hu", "Yurun Chen", "Shengyu Zhang", "Hongxia Yang", "Fan Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8f5fdf474c1cb0d7e9f42d617de27afae38a9fd8f5e0b07d1c86e082dc872a35", "submitted_at": "2025-05-08T17:31:20Z", "updated_at": "2025-11-17T09:12:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.05440v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2505.05440", "identifiers": {"arxiv": "2505.05440"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.05440"}], "name": "EcoAgent: An Efficient Device-Cloud Collaborative Multi-Agent Framework for Mobile Automation", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-05-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "EcoAgent splits mobile automation between cloud planning and on-device observation and verification. Compact screen descriptions and feedback reduce communication costs, with evaluation on AndroidWorld.", "topics": ["mobile-navigation", "planning", "efficiency"], "updated_at": "2025-11-17", "url": "https://arxiv.org/abs/2505.05440"}
{"aliases": [], "authors": ["Zhun Wang", "Vincent Siu", "Zhe Ye", "Tianneng Shi", "Yuzhou Nie", "Xuandong Zhao", "Chenguang Wang", "Wenbo Guo", "Dawn Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eabcb9e11b41d2b2ca7d167f925effc243976a1215d6cb3ea42bb499dd6cce9c", "submitted_at": "2025-05-09T07:40:17Z", "updated_at": "2025-06-14T01:50:21Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.05849v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.05849", "identifiers": {"arxiv": "2505.05849"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.05849"}], "name": "AgentVigil: Generic Black-Box Red-teaming for Indirect Prompt Injection against LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-14", "url": "https://arxiv.org/abs/2505.05849"}
{"aliases": [], "authors": ["Haoyang Xie", "Feng Ju"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9d9534afc30289c04e048295a0bbce5b01279ac56a5e0e418963f13d2842c06c", "submitted_at": "2025-05-10T04:47:08Z", "updated_at": "2025-05-10T04:47:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.06507v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.06507", "identifiers": {"arxiv": "2505.06507"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.06507"}], "name": "Text-to-CadQuery: A New Paradigm for CAD Generation with Scalable Large Model Capabilities", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-10", "url": "https://arxiv.org/abs/2505.06507"}
{"aliases": [], "authors": ["Peng Li", "Suizhi Ma", "Jialiang Chen", "Yuan Liu", "Congyi Zhang", "Wei Xue", "Wenhan Luo", "Alla Sheffer", "Wenping Wang", "Yike Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e9e7da0f8b1537b4e604065a6381fe16ba22565a18d685cf83fc98a57f3bff97", "submitted_at": "2025-05-11T14:54:26Z", "updated_at": "2025-07-06T12:36:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.07003v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.07003", "identifiers": {"arxiv": "2505.07003"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.07003"}], "name": "CMD: Controllable Multiview Diffusion for 3D Editing and Progressive Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-06", "url": "https://arxiv.org/abs/2505.07003"}
{"aliases": [], "authors": ["Kai Xu", "YiWei Mao", "XinYi Guan", "ZiLong Feng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eb27fc805cf0b88e0af8d5ead033a627cfc6ab2f7848be8e13b6c721851fd192", "submitted_at": "2025-05-12T12:06:23Z", "updated_at": "2025-05-12T12:06:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.07473v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.07473", "identifiers": {"arxiv": "2505.07473"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.07473"}], "name": "Web-Bench: A LLM Code Benchmark Based on Web Standards and Frameworks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-12", "url": "https://arxiv.org/abs/2505.07473"}
{"aliases": [], "authors": ["Shuning Zhang", "Jingruo Chen", "Zhiqi Gao", "Jiajing Gao", "Xin Yi", "Hewu Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63f4a505645fec16eac6b59923a6c1979c53376f19c6a4ea53c9a5d32a45e3d4", "submitted_at": "2025-05-15T00:42:51Z", "updated_at": "2025-05-16T07:01:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.09875v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.09875", "identifiers": {"arxiv": "2505.09875"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.09875"}], "name": "Characterizing Unintended Consequences in Human-GUI Agent Collaboration for Web Browsing", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-05-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies unintended consequences of collaborative web browsing with GUI agents through social-media analysis and interviews. It organizes observed failures, user impacts, and mitigation practices around planning, interface interaction, feedback, privacy, and trust.", "topics": ["human-in-the-loop", "evaluation", "safety-security"], "updated_at": "2025-05-16", "url": "https://arxiv.org/abs/2505.09875"}
{"aliases": [], "authors": ["Ranjan Sapkota", "Konstantinos I. Roumeliotis", "Manoj Karkee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "888b0796c02f332e87fb8589cb0639f29ad7e3b8bd89ede1f74367aaaaa9c236", "submitted_at": "2025-05-15T16:21:33Z", "updated_at": "2025-09-30T04:21:32Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2505.10468v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.10468", "identifiers": {"arxiv": "2505.10468"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.10468"}], "name": "AI Agents vs. Agentic AI: A Conceptual Taxonomy, Applications and Challenges", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-30", "url": "https://arxiv.org/abs/2505.10468"}
{"aliases": [], "authors": ["Shanhui Zhao", "Hao Wen", "Wenjie Du", "Cheng Liang", "Yunxin Liu", "Xiaozhou Ye", "Ye Ouyang", "Yuanchun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08cc5be720ac4e82560ad6b7549f4c38c24603d23d9d051894d1078c247059e2", "submitted_at": "2025-05-15T05:28:35Z", "updated_at": "2025-05-15T05:28:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.10593v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2505.10593", "identifiers": {"arxiv": "2505.10593"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.10593"}], "name": "LLM-Explorer: Towards Efficient and Affordable LLM-based Exploration for Mobile Apps", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-05-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "LLM-Explorer uses language models to maintain compact knowledge for mobile app exploration, then generates actions without invoking the model at every step. The study compares exploration coverage and cost across twenty apps.", "topics": ["mobile-navigation", "efficiency"], "updated_at": "2025-05-15", "url": "https://arxiv.org/abs/2505.10593"}
{"aliases": [], "authors": ["Omar Shaikh", "Shardul Sapkota", "Shan Rizvi", "Eric Horvitz", "Joon Sung Park", "Diyi Yang", "Michael S. Bernstein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d986ee6aa62b11759c366fa9f6a01a1a0b12eae715970eacd8053dbbb6622fe5", "submitted_at": "2025-05-16T04:00:31Z", "updated_at": "2025-09-21T00:03:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.10831v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.10831", "identifiers": {"arxiv": "2505.10831"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.10831"}], "name": "Creating General User Models from Computer Use", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-05-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes general user models that turn computer-use observations into confidence-weighted propositions and revise them as new context arrives. Demonstrations use this memory to contextualize assistants, manage notifications, and support proactive actions across apps.", "topics": ["memory", "human-in-the-loop", "cross-app"], "updated_at": "2025-09-21", "url": "https://arxiv.org/abs/2505.10831"}
{"aliases": [], "authors": ["Bin Lei", "Weitai Kang", "Zijian Zhang", "Winson Chen", "Xi Xie", "Shan Zuo", "Mimi Xie", "Ali Payani", "Mingyi Hong", "Yan Yan", "Caiwen Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e5e0ff8846bd74cf0ba92f1ef65ef0309967274d7da347c62cda2940ceba25a", "submitted_at": "2025-05-16T05:43:27Z", "updated_at": "2026-05-01T07:44:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.10887v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.10887", "identifiers": {"arxiv": "2505.10887"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.10887"}], "name": "InfantAgent-Next: A Multimodal Generalist Agent for Automated Computer Interaction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-01", "url": "https://arxiv.org/abs/2505.10887"}
{"aliases": [], "authors": ["Ada Chen", "Yongjiang Wu", "Junyuan Zhang", "Jingyu Xiao", "Shu Yang", "Jen-tse Huang", "Kun Wang", "Wenxuan Wang", "Shuai Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8a0890cd57809ff5e3bede6dc2b851b42fad9307a5bfa0ddd744cc2ac70d0c2", "submitted_at": "2025-05-16T06:56:42Z", "updated_at": "2026-04-29T05:56:22Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.10924v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.10924", "identifiers": {"arxiv": "2505.10924"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.10924"}], "name": "A Survey on the Safety and Security Threats of Computer-Using Agents: JARVIS or Ultron?", "notes": "", "published_at": "2025-05-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Systematizes safety and security threats to computer-using agents, with taxonomies of threats and defensive strategies.", "topics": ["survey", "safety-security"], "updated_at": "2026-04-29", "url": "https://arxiv.org/abs/2505.10924"}
{"aliases": [], "authors": ["Xilong Wang", "John Bloch", "Zedian Shao", "Yuepeng Hu", "Shuyan Zhou", "Neil Zhenqiang Gong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "989673e9cb9103f8c7a28a0b4d838bd83c6e5d87b7f5e44bdb6753d7f0c69a88", "submitted_at": "2025-05-16T22:00:26Z", "updated_at": "2025-10-17T01:52:39Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.11717v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.11717", "identifiers": {"arxiv": "2505.11717"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.11717"}], "name": "WebInject: Prompt Injection Attack to Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-17", "url": "https://arxiv.org/abs/2505.11717"}
{"aliases": [], "authors": ["Weikai Xu", "Zhizheng Jiang", "Yuxuan Liu", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Yuanchun Li", "Yunxin Liu", "Bin Wang", "Bo An"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5ca1ebfdd5a96e146d37539294462c7fb67bf20deaf77307de37aa3bb1dde082", "submitted_at": "2025-05-17T07:58:34Z", "updated_at": "2026-02-02T06:55:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.11891v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2505.11891", "identifiers": {"arxiv": "2505.11891"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.11891"}], "name": "Mobile-Bench-v2: A More Realistic and Comprehensive Benchmark for VLM-based Mobile Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Mobile-Bench-v2 evaluates mobile GUI agents using multiple acceptable offline action paths, noisy interfaces with ads or pop-ups, and ambiguous instructions with scripted clarification exchanges. These splits separate step-level execution, robustness, and proactive interaction.", "topics": ["evaluation", "mobile-navigation", "robustness", "human-in-the-loop"], "updated_at": "2026-02-02", "url": "https://arxiv.org/abs/2505.11891"}
{"aliases": [], "authors": ["Tianxiong Zhong", "Xingye Tian", "Boyuan Jiang", "Xuebo Wang", "Xin Tao", "Pengfei Wan", "Zhiwei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c93f1673c9bb32d40bd2f11fb2dbee9ec7c8246741c589cb3142e84cdb733fe9", "submitted_at": "2025-05-17T15:32:54Z", "updated_at": "2025-09-28T09:51:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.12053v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.12053", "identifiers": {"arxiv": "2505.12053"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.12053"}], "name": "VFRTok: Variable Frame Rates Video Tokenizer with Duration-Proportional Information Assumption", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-28", "url": "https://arxiv.org/abs/2505.12053"}
{"aliases": [], "authors": ["Kun Huang", "Weikai Xu", "Yuxuan Liu", "Quandong Wang", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Bin Wang", "Bo An"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9bc92029e16b5afb4b54837d2ab2304c7b8176548061c9deafdca101754128dc", "submitted_at": "2025-05-18T08:28:05Z", "updated_at": "2026-03-21T09:25:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.12299v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2505.12299", "identifiers": {"arxiv": "2505.12299"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12299"}], "name": "MobileIPL: Enhancing Mobile Agents Thinking Process via Iterative Preference Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileIPL samples trees of action-planning thoughts, propagates rule-based leaf rewards backward, and forms preference pairs for reasoning-level optimization. It also expands screenshot-based instruction data through staged evolution before evaluation on mobile GUI-agent benchmarks.", "topics": ["mobile-navigation", "reinforcement-learning", "planning", "data-generation"], "updated_at": "2026-03-21", "url": "https://arxiv.org/abs/2505.12299"}
{"aliases": [], "authors": ["Xinbin Yuan", "Jian Zhang", "Kaixin Li", "Zhuoxuan Cai", "Lujian Yao", "Jie Chen", "Enguang Wang", "Qibin Hou", "Jinwei Chen", "Peng-Tao Jiang", "Bo Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755f885b79e9f3eef9b4cbdb8593fb3cde1ac8a92f86df52d5d4203eb6cad039", "submitted_at": "2025-05-18T11:22:04Z", "updated_at": "2025-05-24T03:01:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.12370v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.12370", "identifiers": {"arxiv": "2505.12370"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12370"}], "name": "Enhancing Visual Grounding for GUI Agents via Self-Evolutionary Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes GUI grounding training that combines curated seed examples, continuous location-dependent policy feedback, and iterative refinement guided by attention maps. The study evaluates the resulting reinforcement-learning procedure on three grounding benchmarks, including ScreenSpot-Pro.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-05-24", "url": "https://arxiv.org/abs/2505.12370"}
{"aliases": [], "authors": ["Yedidel Louck", "Ariel Stulman", "Amit Dvir"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac8a925229769ffb0fe0509856805a46015eeab4cb9bbdfc731d1af77e7f3bb3", "submitted_at": "2025-05-18T16:25:21Z", "updated_at": "2025-08-28T20:13:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.12490v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.12490", "identifiers": {"arxiv": "2505.12490"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.12490"}], "name": "Improving Google A2A Protocol: Protecting Sensitive Data and Mitigating Unintended Harms in Multi-Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-28", "url": "https://arxiv.org/abs/2505.12490"}
{"aliases": [], "authors": ["Longxi Gao", "Li Zhang", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Mengwei Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "10d826ea7715ab7054522a9dcba622d5dae3e6c3faee421e4b168e16104d184b", "submitted_at": "2025-05-18T16:34:30Z", "updated_at": "2025-10-10T09:28:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.12493v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.12493", "identifiers": {"arxiv": "2505.12493"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12493"}], "name": "GUI-Shift: Enhancing VLM-Based GUI Agents through Self-supervised Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Shift trains an inverse-dynamics task that predicts the initial action connecting two GUI states, using trajectories without natural-language instructions. It combines rule-based reinforcement learning with data filtering and evaluates transfer to Android automation and screenshot grounding.", "topics": ["reinforcement-learning", "data-generation", "grounding", "mobile-navigation"], "updated_at": "2025-10-10", "url": "https://arxiv.org/abs/2505.12493"}
{"aliases": [], "authors": ["Yunseok Jang", "Yeda Song", "Sungryull Sohn", "Lajanugen Logeswaran", "Tiange Luo", "Dong-Ki Kim", "Kyunghoon Bae", "Honglak Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64672e6a3450256a8b04d9db231e885e2dd4e8688235f6c7853d70d6276610ff", "submitted_at": "2025-05-19T02:39:03Z", "updated_at": "2025-05-19T02:39:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.12632v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2505.12632", "identifiers": {"arxiv": "2505.12632"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12632"}], "name": "Scalable Video-to-Dataset Generation for Cross-Platform Mobile Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MONDAY and an automated pipeline that extracts mobile-navigation training data from instructional videos across operating-system platforms.", "topics": ["video", "data-generation", "mobile-navigation"], "updated_at": "2025-05-19", "url": "https://arxiv.org/abs/2505.12632"}
{"aliases": [], "authors": ["Zihua Wang", "Ruibo Li", "Haozhe Du", "Joey Tianyi Zhou", "Yu Zhang", "Xu Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5987a7a62c584df5ffdf13ee8d32f3e679e84d40d74f987c671f8c97f51afe4d", "submitted_at": "2025-05-19T05:35:30Z", "updated_at": "2026-02-03T12:51:29Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.12728v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.12728", "identifiers": {"arxiv": "2505.12728"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.12728"}], "name": "SpecFLASH: A Latent-Guided Semi-autoregressive Speculative Decoding Framework for Efficient Multimodal Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-03", "url": "https://arxiv.org/abs/2505.12728"}
{"aliases": [], "authors": ["Zheng Wu", "Pengzhou Cheng", "Zongru Wu", "Lingzhong Dong", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce9570eb6a52feb220cdd2eb4c7f37a2ea63082febe48888a2654c974138a82d", "submitted_at": "2025-05-19T08:29:05Z", "updated_at": "2025-11-22T15:07:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.12842v4"}}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.12842", "identifiers": {"arxiv": "2505.12842"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12842"}], "name": "GEM: Gaussian Embedding Modeling for Out-of-Distribution Detection in GUI Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-05-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "GEM detects instructions outside a GUI agent’s capabilities by modeling distances in its input-embedding space with a Gaussian mixture. Experiments across mobile, computer and browser data also examine selective escalation to a cloud model when an input is detected as unfamiliar.", "topics": ["robustness", "safety-security", "evaluation", "efficiency"], "updated_at": "2025-11-22", "url": "https://arxiv.org/abs/2505.12842"}
{"aliases": [], "authors": ["Liangxuan Wu", "Chao Wang", "Tianming Liu", "Yanjie Zhao", "Haoyu Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9657a66aebe46151ed56bf76131ffe45f52459c1cf0e000907113776c91259d9", "submitted_at": "2025-05-19T11:17:46Z", "updated_at": "2025-05-20T07:02:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.12981v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2505.12981", "identifiers": {"arxiv": "2505.12981"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12981"}], "name": "From Assistants to Adversaries: Exploring the Security Risks of Mobile LLM Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes security vulnerabilities across OEM system-level, third-party, and research mobile LLM agents, covering GUI and cross-application attack surfaces.", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2025-05-20", "url": "https://arxiv.org/abs/2505.12981"}
{"aliases": [], "authors": ["Mykyta Mudryi", "Markiyan Chaklosh", "Grzegorz Wójcik"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "709ee511ffa39e805fcc58abe066cf2cec6c1707a4a9c64d9cfd107bef7544af", "submitted_at": "2025-05-19T13:10:29Z", "updated_at": "2025-05-19T13:10:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.13076v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2505.13076", "identifiers": {"arxiv": "2505.13076"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.13076"}], "name": "The Hidden Dangers of Browsing AI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Outlines an end-to-end threat model and empirical security evaluation across architectural layers of autonomous browser agents.", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2025-05-19", "url": "https://arxiv.org/abs/2505.13076"}
{"aliases": [], "authors": ["Tianbao Xie", "Jiaqi Deng", "Xiaochuan Li", "Junlin Yang", "Haoyuan Wu", "Jixuan Chen", "Wenjing Hu", "Xinyuan Wang", "Yuhui Xu", "Zekun Wang", "Yiheng Xu", "Junli Wang", "Doyen Sahoo", "Tao Yu", "Caiming Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bec4ad578c01ebdfb58bd269562d3ca7a2bd8d2687cde07cb867db0b78cb105", "submitted_at": "2025-05-19T15:09:23Z", "updated_at": "2025-10-24T19:08:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.13227v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.13227", "identifiers": {"arxiv": "2505.13227"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.13227"}], "name": "Scaling Computer-Use Grounding via User Interface Decomposition and Synthesis", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-05-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces OSWorld-G grounding evaluation and the Jedi synthesis pipeline for examples spanning layout reasoning and fine-grained interaction.", "topics": ["grounding", "data-generation", "evaluation"], "updated_at": "2025-10-24", "url": "https://arxiv.org/abs/2505.13227"}
{"aliases": [], "authors": ["Yifu Cai", "Xinyu Li", "Mononito Goswami", "Michał Wiliński", "Gus Welter", "Artur Dubrawski"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ad64dec5d6a0e78171a387d67b072e262f88a2d3f5b6724cdcd1c6ac54ad6f63", "submitted_at": "2025-05-19T16:11:23Z", "updated_at": "2025-05-19T16:11:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.13291v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.13291", "identifiers": {"arxiv": "2505.13291"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.13291"}], "name": "TimeSeriesGym: A Scalable Benchmark for (Time Series) Machine Learning Engineering Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-19", "url": "https://arxiv.org/abs/2505.13291"}
{"aliases": [], "authors": ["Junyang Wang", "Haiyang Xu", "Xi Zhang", "Ming Yan", "Ji Zhang", "Fei Huang", "Jitao Sang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d723cb764deb78b194ff0199c7f1b921a12b6a41655c07b06ae88c3d0af58358", "submitted_at": "2025-05-20T03:48:19Z", "updated_at": "2025-06-03T08:39:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.13887v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["identifiers.preferred_arxiv", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2505.13887", "identifiers": {"arxiv": "2505.13887", "preferred_arxiv": "2502.17110"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.13887"}], "name": "Mobile-Agent-V: A Video-Guided Approach for Effortless and Efficient Operational Knowledge Injection in Mobile Automation", "notes": "The authors identify this arXiv entry as an accidental duplicate submission and direct future updates to https://arxiv.org/abs/2502.17110. Retained as a bibliographic provenance record; not counted as an additional reviewed work.", "published_at": "2025-05-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-03", "url": "https://arxiv.org/abs/2505.13887"}
{"aliases": [], "authors": ["Yanheng He", "Jiahe Jin", "Pengfei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "026a64046926912a54e6a7c6cf62a17740a6f89a94713615f9750f8bf8302c78", "submitted_at": "2025-05-20T04:20:18Z", "updated_at": "2026-03-03T03:55:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.13909v2"}}, "environments": ["windows", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.13909", "identifiers": {"arxiv": "2505.13909"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.13909"}], "name": "Efficient Agent Training for Computer Use", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "PC Agent-E augments 312 human computer-use trajectories with alternative action decisions synthesized by Claude 3.7 Sonnet. It studies training with this combination and introduces WindowsAgentArena-V2 as an updated evaluation benchmark.", "topics": ["data-generation", "evaluation"], "updated_at": "2026-03-03", "url": "https://arxiv.org/abs/2505.13909"}
{"aliases": [], "authors": ["Zhiqian Lan", "Yuxuan Jiang", "Ruiqi Wang", "Xuanbing Xie", "Rongkui Zhang", "Yicheng Zhu", "Peihang Li", "Tianshuo Yang", "Tianxing Chen", "Haoyu Gao", "Xiaokang Yang", "Xuelong Li", "Hongyuan Zhang", "Yao Mu", "Ping Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53dc342e2afbff2d56ca78bd0cbb581668f6c7428700399bc52efa5cdfa7333b", "submitted_at": "2025-05-20T07:29:26Z", "updated_at": "2025-05-29T01:50:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.14030v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.14030", "identifiers": {"arxiv": "2505.14030"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.14030"}], "name": "AutoBio: A Simulation and Benchmark for Robotic Automation in Digital Biology Laboratory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-29", "url": "https://arxiv.org/abs/2505.14030"}
{"aliases": [], "authors": ["Fanglin Mo", "Junzhe Chen", "Haoxuan Zhu", "Xuming Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de6c501a42c3828bdf795ac59c38b982e738c38bd5d0c59cfb6cb10b51e39292", "submitted_at": "2025-05-20T09:45:55Z", "updated_at": "2025-05-20T09:45:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.14141v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.14141", "identifiers": {"arxiv": "2505.14141"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.14141"}], "name": "Building a Stable Planner: An Extended Finite State Machine Based Planning Module for Mobile GUI Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "SPlanner models mobile-app functions and configurations with extended finite-state machines, traverses them to construct execution paths, and converts those paths into natural-language plans. A VLM follows these plans to perform GUI actions, with evaluation on AndroidWorld.", "topics": ["planning", "mobile-navigation"], "updated_at": "2025-05-20", "url": "https://arxiv.org/abs/2505.14141"}
{"aliases": [], "authors": ["Yijie Lu", "Manman Zhao", "Tianjie Ju", "Zihe Yan", "Xinbei Ma", "Yuan Guo", "Daizong Ding", "Gongshen Liu", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c256d8b0d774efdeee896860f8c504c0d1593215508fdf0881d3fd3cdc3968c", "submitted_at": "2025-05-20T12:41:05Z", "updated_at": "2026-06-05T10:24:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.14289v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.14289", "identifiers": {"arxiv": "2505.14289"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.14289"}], "name": "EVA: Evolving Semantic Adversaries for Red-Teaming GUI Agents Against Environmental Injection Attacks", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-05-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces EVA, a red-teaming framework that evolves semantic payloads for GUI environmental-injection tests. It studies the contribution of semantic versus visual deception and distills discovered patterns into reusable evaluation rules.", "topics": ["safety-security", "evaluation"], "updated_at": "2026-06-05", "url": "https://arxiv.org/abs/2505.14289"}
{"aliases": [], "authors": ["Pengzhou Cheng", "Haowen Hu", "Zheng Wu", "Zongru Wu", "Tianjie Ju", "Zhuosheng Zhang", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "815345ed44d70f40ce9784c212d269ffaaa476e5c2ea5cdbbaaeb98d22eba3cb", "submitted_at": "2025-05-20T14:29:18Z", "updated_at": "2026-06-15T06:42:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.14418v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.14418", "identifiers": {"arxiv": "2505.14418"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.14418"}], "name": "Hidden Ghost Hand: Unveiling Backdoor Vulnerabilities in MLLM-Powered Mobile GUI Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-05-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies backdoor risks in mobile GUI agents through AgentGhost, which combines goal-level and interaction-history triggers during model training. The paper evaluates multiple mobile agents and a tailored defense.", "topics": ["safety-security", "mobile-navigation", "interaction-history"], "updated_at": "2026-06-15", "url": "https://arxiv.org/abs/2505.14418"}
{"aliases": [], "authors": ["Xuyang Liu", "Yiyu Wang", "Junpeng Ma", "Linfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d00f8c91c1c80cddffcecec37764d1fcaa8233a5ba25a753c585fc9998f401c", "submitted_at": "2025-05-20T14:52:31Z", "updated_at": "2025-11-18T08:19:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.14454v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.14454", "identifiers": {"arxiv": "2505.14454"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.14454"}], "name": "Video Compression Commander: Plug-and-Play Inference Acceleration for Video Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-18", "url": "https://arxiv.org/abs/2505.14454"}
{"aliases": [], "authors": ["Anna C. Doris", "Md Ferdous Alam", "Amin Heyrani Nobari", "Faez Ahmed"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "10481fd9f4f28eb6d7d1e5eb692a95981651570e0be4c57ad4bcadedfb8c57e0", "submitted_at": "2025-05-20T17:34:44Z", "updated_at": "2025-05-20T17:34:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.14646v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.14646", "identifiers": {"arxiv": "2505.14646"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.14646"}], "name": "CAD-Coder: An Open-Source Vision-Language Model for Computer-Aided Design Code Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-20", "url": "https://arxiv.org/abs/2505.14646"}
{"aliases": [], "authors": ["Shan Chen", "Pedro Moreira", "Yuxin Xiao", "Sam Schmidgall", "Jeremy Warner", "Hugo Aerts", "Thomas Hartvigsen", "Jack Gallifant", "Danielle S. Bitterman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a227e1d3ab5d05be9df0de9df0de1ae4842c2757ba522304feca3b6008473661", "submitted_at": "2025-05-20T22:42:33Z", "updated_at": "2025-05-20T22:42:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.14963v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.14963", "identifiers": {"arxiv": "2505.14963"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.14963"}], "name": "MedBrowseComp: Benchmarking Medical Deep Research and Computer Use", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-20", "url": "https://arxiv.org/abs/2505.14963"}
{"aliases": [], "authors": ["Hyunseok Lee", "Jeonghoon Kim", "Beomjun Kim", "Jihoon Tack", "Chansong Jo", "Jaehong Lee", "Cheonbok Park", "Sookyo In", "Jinwoo Shin", "Kang Min Yoo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9acb7a88a0cf07427c1878c896dec920852810879c30bad8692358f2e2016539", "submitted_at": "2025-05-21T08:36:18Z", "updated_at": "2025-05-24T15:08:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15259v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.15259", "identifiers": {"arxiv": "2505.15259"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.15259"}], "name": "ReGUIDE: Data Efficient GUI Grounding via Spatial Reasoning and Search", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines self-generated localization reasoning, spatial consistency checks, and inference-time search to improve data-efficient web grounding.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-05-24", "url": "https://arxiv.org/abs/2505.15259"}
{"aliases": [], "authors": ["Hyungjoo Chae", "Sunghwan Kim", "Junhee Cho", "Seungone Kim", "Seungjun Moon", "Gyeom Hwangbo", "Dongha Lim", "Minjin Kim", "Yeonjun Hwang", "Minju Gwak", "Dongwook Choi", "Minseok Kang", "Gwanhoon Im", "ByeongUng Cho", "Hyojun Kim", "Jun Hee Han", "Taeyoon Kwon", "Minju Kim", "Beong-woo Kwak", "Dongjin Kang", "Jinyoung Yeo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "440abdc75c0a51478cfedc6ee0a9196ccff641ec5a23849ba56be43448bab15f", "submitted_at": "2025-05-21T08:56:55Z", "updated_at": "2025-11-25T21:34:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15277v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.15277", "identifiers": {"arxiv": "2505.15277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.15277"}], "name": "Web-Shepherd: Advancing PRMs for Reinforcing Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Web-Shepherd is a process reward model for assessing individual steps of web-navigation trajectories. It introduces a collection of 40,000 step-level preference pairs with checklists and a separate benchmark for comparing reward models, then evaluates verification during WebArena-lite navigation.", "topics": ["reinforcement-learning", "evaluation", "data-generation", "web-navigation"], "updated_at": "2025-11-25", "url": "https://arxiv.org/abs/2505.15277"}
{"aliases": [], "authors": ["Yuchen Yan", "Jin Jiang", "Zhenbang Ren", "Yijun Li", "Xudong Cai", "Yang Liu", "Xin Xu", "Mengdi Zhang", "Jian Shao", "Yongliang Shen", "Jun Xiao", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d726e26526a2b18bb805cb816500f9733f9dba0e1457536efac0785cea5cdc9d", "submitted_at": "2025-05-21T17:54:43Z", "updated_at": "2026-02-18T08:45:15Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.15801v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.15801", "identifiers": {"arxiv": "2505.15801"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.15801"}], "name": "VerifyBench: Benchmarking Reference-based Reward Systems for Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-18", "url": "https://arxiv.org/abs/2505.15801"}
{"aliases": [], "authors": ["Yuqi Zhou", "Sunhao Dai", "Shuai Wang", "Kaiwen Zhou", "Qinglin Jia", "Jun Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "321f2d627451d60e130c17574994648fae69fc784b0ed03f6560e7f1ef9eb023", "submitted_at": "2025-05-21T17:59:09Z", "updated_at": "2025-05-22T11:15:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15810v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.15810", "identifiers": {"arxiv": "2505.15810"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.15810"}], "name": "GUI-G1: Understanding R1-Zero-Like Training for Visual Grounding in GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes reasoning templates, reward design, and policy updates in GUI grounding, then proposes targeted adjustments to the training process.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-05-22", "url": "https://arxiv.org/abs/2505.15810"}
{"aliases": [], "authors": ["Penghao Wu", "Lewei Lu", "Ziwei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "396bffe767e4f9dd539402cf644dce462f3a0dda8a003ca64e2a8515a8d0ae9c", "submitted_at": "2025-05-21T17:59:52Z", "updated_at": "2025-05-21T17:59:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.15816v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.15816", "identifiers": {"arxiv": "2505.15816"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.15816"}], "name": "Streamline Without Sacrifice -- Squeeze out Computation Redundancy in LMM", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-21", "url": "https://arxiv.org/abs/2505.15816"}
{"aliases": [], "authors": ["Tianyi Ma", "Yiyue Qian", "Zheyuan Zhang", "Zehong Wang", "Xiaoye Qian", "Feifan Bai", "Yifan Ding", "Xuwei Luo", "Shinan Zhang", "Keerthiram Murugesan", "Chuxu Zhang", "Yanfang Ye"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f83130cb80d73287ecbc2a548ba0422c35adec45541f22e76c4ee5b1750d7a7f", "submitted_at": "2025-05-21T04:32:35Z", "updated_at": "2025-05-21T04:32:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.15859v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2505.15859", "identifiers": {"arxiv": "2505.15859"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.15859"}], "name": "AutoData: A Multi-Agent System for Open Web Data Collection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-21", "url": "https://arxiv.org/abs/2505.15859"}
{"aliases": [], "authors": ["Fanbin Lu", "Zhisheng Zhong", "Shu Liu", "Chi-Wing Fu", "Jiaya Jia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c092569d4d79026ddbdcc3bc9da1cfc617ff44886dab0b36f311cc81083e0fb5", "submitted_at": "2025-05-22T06:24:32Z", "updated_at": "2025-05-22T06:24:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.16282v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.16282", "identifiers": {"arxiv": "2505.16282"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.16282"}], "name": "ARPO:End-to-End Policy Optimization for GUI Agents with Experience Replay", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "ARPO extends group-relative policy optimization with a replay buffer that reuses successful GUI trajectories across training iterations. It filters tasks using baseline performance and compares end-to-end reinforcement learning with offline preference optimization on OSWorld.", "topics": ["reinforcement-learning", "long-horizon"], "updated_at": "2025-05-22", "url": "https://arxiv.org/abs/2505.16282"}
{"aliases": [], "authors": ["Zhepei Wei", "Wenlin Yao", "Yao Liu", "Weizhi Zhang", "Qin Lu", "Liang Qiu", "Changlong Yu", "Puyang Xu", "Chao Zhang", "Bing Yin", "Hyokun Yun", "Lihong Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f311fb50b36c234787e652b5689246460153a8f442e83b312048cdf6a4fb1502", "submitted_at": "2025-05-22T09:07:43Z", "updated_at": "2025-10-08T21:16:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.16421v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.16421", "identifiers": {"arxiv": "2505.16421"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.16421"}], "name": "WebAgent-R1: Training Web Agents via End-to-End Multi-Turn Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains web agents through asynchronous multi-turn environment interaction, using task-success rewards to optimize complete trajectories.", "topics": ["reinforcement-learning", "web-navigation"], "updated_at": "2025-10-08", "url": "https://arxiv.org/abs/2505.16421"}
{"aliases": [], "authors": ["Xiaoran Yin", "Xu Luo", "Hao Wu", "Lianli Gao", "Jingkuan Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "93c67222fb0f581e9f9cb3ff59c8d28faec9dc037180060ad5ac962b4469144c", "submitted_at": "2025-05-22T09:08:47Z", "updated_at": "2025-09-05T02:24:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.16422v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2505.16422", "identifiers": {"arxiv": "2505.16422"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.16422"}], "name": "Unlocking Smarter Device Control: Foresighted Planning with a World Model-Driven Code Execution Approach", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-05-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "FPWC builds and refines a task-oriented world model before planning mobile-device actions. It executes plans as code and evaluates the approach in simulated environments and on real mobile devices.", "topics": ["world-models", "planning", "mobile-navigation"], "updated_at": "2025-09-05", "url": "https://arxiv.org/abs/2505.16422"}
{"aliases": [], "authors": ["Bin Xie", "Rui Shao", "Gongwei Chen", "Kaiwen Zhou", "Yinchuan Li", "Jie Liu", "Min Zhang", "Liqiang Nie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "641001c49fccac1e1c0659fe765eba5a30cefd2a7ff5d00c821f79092c45a1df", "submitted_at": "2025-05-22T16:01:06Z", "updated_at": "2025-05-22T16:01:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.16827v1"}, "publication": {"authors": ["Bin Xie", "Rui Shao", "Gongwei Chen", "Kaiwen Zhou", "Yinchuan Li", "Jie Liu", "Min Zhang", "Liqiang Nie"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.282", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.282/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "5650-5667", "pdf_url": "https://aclanthology.org/2025.acl-long.282.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "GUI-explorer: Autonomous Exploration and Mining of Transition-aware Knowledge for GUI Agent", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.282/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:59Z", "role": "primary", "sha256": "40babce7d310777fa65a498c170320124d4d4941f2cb46ad0ff04aae3583e02f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.282/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.16827", "identifiers": {"arxiv": "2505.16827"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.16827"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.282/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.282.pdf"}], "name": "GUI-explorer: Autonomous Exploration and Mining of Transition-aware Knowledge for GUI Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-explorer generates exploration goals from screenshots and interface structure, then mines observation-action-outcome triples for app-specific transition knowledge. The resulting training-free agent uses this knowledge during mobile task execution, evaluated on SPA-Bench and AndroidWorld.", "topics": ["mobile-navigation", "data-generation", "memory", "planning"], "updated_at": "2025-05-22", "url": "https://arxiv.org/abs/2505.16827"}
{"aliases": [], "authors": ["Shilin Yan", "Jiaming Han", "Joey Tsai", "Hongwei Xue", "Rongyao Fang", "Lingyi Hong", "Ziyu Guo", "Ray Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "10a728cf4999ebf4661b4e14cfa2c653030202831fd374187b18cfca8cbd6b18", "submitted_at": "2025-05-22T17:59:53Z", "updated_at": "2025-12-21T07:08:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.17020v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.17020", "identifiers": {"arxiv": "2505.17020"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.17020"}], "name": "CrossLMM: Decoupling Long Video Sequences from LMMs via Dual Cross-Attention Mechanisms", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-21", "url": "https://arxiv.org/abs/2505.17020"}
{"aliases": [], "authors": ["Xueyang Zhou", "Weidong Wang", "Lin Lu", "Jiawen Shi", "Guiyao Tie", "Yongtian Xu", "Lixing Chen", "Pan Zhou", "Neil Zhenqiang Gong", "Lichao Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c2cc928e332b21b823f37b53979d0cf964618098e65052d7504a4375dd69128", "submitted_at": "2025-05-23T10:56:06Z", "updated_at": "2025-07-18T07:34:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.17735v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2505.17735", "identifiers": {"arxiv": "2505.17735"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.17735"}], "name": "SafeAgent: Safeguarding LLM Agents via an Automated Risk Simulator", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-18", "url": "https://arxiv.org/abs/2505.17735"}
{"aliases": [], "authors": ["Danyang Zhang", "Situo Zhang", "Ziyue Yang", "Zichen Zhu", "Zihan Zhao", "Ruisheng Cao", "Lu Chen", "Kai Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40cbdc37216c58acd1b7523c48ad735df9bde485405a4e7b439aec85730c5bc0", "submitted_at": "2025-05-23T17:23:11Z", "updated_at": "2025-05-23T17:23:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.18121v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.18121", "identifiers": {"arxiv": "2505.18121"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.18121"}], "name": "ProgRM: Build Better GUI Agents with Progress Rewards", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "ProgRM predicts per-step task completion progress to provide intermediate rewards during online GUI-agent training. It derives progress labels by using longest common subsequences to identify key trajectory steps, allowing useful partial progress to contribute even when a trajectory ultimately fails.", "topics": ["reinforcement-learning", "data-generation"], "updated_at": "2025-05-23", "url": "https://arxiv.org/abs/2505.18121"}
{"aliases": [], "authors": ["Peijie Yu", "Yifan Yang", "Jinjian Li", "Zelong Zhang", "Haorui Wang", "Xiao Feng", "Feng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1d2129a4e55c04a2d49bf05c375b9599d22a289002438239226316a4d0c1dbf0", "submitted_at": "2025-05-24T15:25:44Z", "updated_at": "2025-06-27T03:58:25Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.18746v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.18746", "identifiers": {"arxiv": "2505.18746"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.18746"}], "name": "$C^3$-Bench: The Things Real Disturbing LLM based Agent in Multi-Tasking", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-27", "url": "https://arxiv.org/abs/2505.18746"}
{"aliases": [], "authors": ["Kai Mei", "Xi Zhu", "Hang Gao", "Shuhang Lin", "Yongfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c382ce428e6dd41ec07af7575461c69906ce386fd2bcd8ad9f0d792d2b8f0b3", "submitted_at": "2025-05-24T18:56:00Z", "updated_at": "2025-10-31T21:48:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.18829v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.18829", "identifiers": {"arxiv": "2505.18829"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.18829"}], "name": "LiteCUA: Computer as MCP Server for Computer-Use Agent on AIOS", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "AIOS 1.0 exposes computer states and actions through an MCP server to provide structured environmental context for agents. LiteCUA is a lightweight agent built on this interface, evaluated on OSWorld to study the effect of computer-environment contextualization.", "topics": ["tool-integration", "automation"], "updated_at": "2025-10-31", "url": "https://arxiv.org/abs/2505.18829"}
{"aliases": [], "authors": ["Runliang Niu", "Jinglong Ji", "Yi Chang", "Qi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb782466eccc7a271c89a4161c7b142f93f8f0d585cad39a1ccad048cf514426", "submitted_at": "2025-05-25T11:13:03Z", "updated_at": "2025-05-25T11:13:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.19095v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:24Z", "role": "primary", "sha256": "645bd2ecbc013dab3cbc601d5a42011d8f7306915ea2ae06baf0767b3962e9ed", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2505.19095v1"}], "id": "arxiv:2505.19095", "identifiers": {"arxiv": "2505.19095"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.19095"}], "name": "ScreenExplorer: Training a Vision-Language Model for Diverse Exploration in Open GUI World", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-05-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains ScreenExplorer with reinforcement learning in dynamic GUI environments, using a world-model curiosity reward and experience-stream distillation to support sustained exploration.", "topics": ["reinforcement-learning", "world-models", "data-generation"], "updated_at": "2025-05-25", "url": "https://arxiv.org/abs/2505.19095"}
{"aliases": [], "authors": ["Yandong Guan", "Xilin Wang", "Ximing Xing", "Jing Zhang", "Dong Xu", "Qian Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a29c1ce5435172e5222a622c37feb0bad94635df9a4d36a292538d65573a53ed", "submitted_at": "2025-05-26T09:01:56Z", "updated_at": "2025-10-21T04:03:20Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.19713v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.19713", "identifiers": {"arxiv": "2505.19713"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.19713"}], "name": "CAD-Coder: Text-to-CAD Generation with Chain-of-Thought and Geometric Reward", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-21", "url": "https://arxiv.org/abs/2505.19713"}
{"aliases": [], "authors": ["Qiushi Sun", "Zhoumianze Liu", "Chang Ma", "Zichen Ding", "Fangzhi Xu", "Zhangyue Yin", "Haiteng Zhao", "Zhenyu Wu", "Kanzhi Cheng", "Zhaoyang Liu", "Jianing Wang", "Qintong Li", "Xiangru Tang", "Tianbao Xie", "Xiachong Feng", "Xiang Li", "Ben Kao", "Wenhai Wang", "Biqing Qi", "Lingpeng Kong", "Zhiyong Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4743f6bfd0ca48717184cf55545e33a229dd5ab2d454385d33ea2f1b25c15b0c", "submitted_at": "2025-05-26T12:27:27Z", "updated_at": "2026-04-20T16:29:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.19897v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2505.19897", "identifiers": {"arxiv": "2505.19897"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.19897"}], "name": "ScienceBoard: Evaluating Multimodal Autonomous Agents in Realistic Scientific Workflows", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "ScienceBoard combines an interactive environment with professional scientific software and 169 human-curated tasks in areas including biochemistry, astronomy, and geoinformatics. It evaluates computer-use agents on visually rich research workflows and analyzes their execution limitations.", "topics": ["evaluation", "science", "workflow"], "updated_at": "2026-04-20", "url": "https://arxiv.org/abs/2505.19897"}
{"aliases": [], "authors": ["Minda Hu", "Tianqing Fang", "Jianshu Zhang", "Junyu Ma", "Zhisong Zhang", "Jingyan Zhou", "Hongming Zhang", "Haitao Mi", "Dong Yu", "Irwin King"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "171071f26a6dc8b9c8851d2f11d60b113983191c905631e431f352711e0d0e26", "submitted_at": "2025-05-26T14:03:37Z", "updated_at": "2025-09-18T11:32:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.20013v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.20013", "identifiers": {"arxiv": "2505.20013"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.20013"}], "name": "WebCoT: Enhancing Web Agent Reasoning by Reconstructing Chain-of-Thought in Reflection, Branching, and Rollback", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-05-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebCoT turns reflection, look-ahead, branching and rollback procedures into chain-of-thought training rationales. Fine-tuning on these reconstructed trajectories is evaluated for web-agent self-improvement and transfer across navigation tasks.", "topics": ["data-generation", "planning", "self-reflection", "web-navigation"], "updated_at": "2025-09-18", "url": "https://arxiv.org/abs/2505.20013"}
{"aliases": [], "authors": ["Yihan Chen", "Benfeng Xu", "Xiaorui Wang", "Yongdong Zhang", "Zhendong Mao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f3c91f5b701650968bb2f3bbb468adf0fabedc476ecb04ba3f1a5206005102be", "submitted_at": "2025-05-26T14:11:12Z", "updated_at": "2025-05-26T14:11:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.20023v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.20023", "identifiers": {"arxiv": "2505.20023"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.20023"}], "name": "Training LLM-Based Agents with Synthetic Self-Reflected Trajectories and Partial Masking", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-26", "url": "https://arxiv.org/abs/2505.20023"}
{"aliases": [], "authors": ["Qinzhuo Wu", "Pengzhi Gao", "Wei Liu", "Jian Luan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2d51388193e19501b580740969a6e08661c6e0a2cf09e2489c46fe4f19209cd", "submitted_at": "2025-05-27T03:09:06Z", "updated_at": "2025-05-27T03:09:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.20660v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.20660", "identifiers": {"arxiv": "2505.20660"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.20660"}], "name": "BacktrackAgent: Enhancing GUI Agent with Error Detection and Backtracking Mechanism", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "BacktrackAgent adds verification, judgment, and reflection modules to detect erroneous GUI actions and guide recovery. It develops backtracking-oriented training data that includes post-action pages and evaluates task completion and step accuracy on Mobile3M and Auto-UI.", "topics": ["robustness", "planning", "data-generation", "mobile-navigation"], "updated_at": "2025-05-27", "url": "https://arxiv.org/abs/2505.20660"}
{"aliases": [], "authors": ["Hanlin Wang", "Chak Tou Leong", "Jiashuo Wang", "Jian Wang", "Wenjie Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08018a5f5d286f25179700b010fe737b3d9c0fe9973236f3ab833f45bf860bb9", "submitted_at": "2025-05-27T05:21:04Z", "updated_at": "2025-05-27T05:21:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.20732v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.20732", "identifiers": {"arxiv": "2505.20732"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.20732"}], "name": "SPA-RL: Reinforcing LLM Agents via Stepwise Progress Attribution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-27", "url": "https://arxiv.org/abs/2505.20732"}
{"aliases": [], "authors": ["Xiang Zhu", "Yichen Liu", "Hezhong Li", "Jianyu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "92b778dd06d54b801fa32210623eb1538f2e337bbdd681642992d8769e83113e", "submitted_at": "2025-05-27T06:56:14Z", "updated_at": "2026-05-29T12:22:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.20795v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.20795", "identifiers": {"arxiv": "2505.20795"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.20795"}], "name": "Learning Generalizable Robot Policy with Human Demonstration Video as a Prompt", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-29", "url": "https://arxiv.org/abs/2505.20795"}
{"aliases": [], "authors": ["Kaiming Liu", "Xuanyu Lei", "Ziyue Wang", "Peng Li", "Yang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4e61bad9172265417e67652ff320808a01aea705456fd1ae9882b4975a3fe403", "submitted_at": "2025-05-27T11:44:50Z", "updated_at": "2026-10-13T06:45:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.21055v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.21055", "identifiers": {"arxiv": "2505.21055"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.21055"}], "name": "Agent-Environment Alignment via Automated Interface Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2505.21055"}
{"aliases": [], "authors": ["Shaoqing Zhang", "Kehai Chen", "Zhuosheng Zhang", "Rumei Li", "Rongxiang Weng", "Yang Xiang", "Min Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4fe40099905712f9d7ba89bc7616d09b260fe6113beb8122cade54a66d46e3d7", "submitted_at": "2025-05-27T14:49:30Z", "updated_at": "2025-09-26T09:13:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.21279v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2505.21279", "identifiers": {"arxiv": "2505.21279"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21279"}], "name": "XBOUND: Exploring Capability Boundaries of Device-Control Agents at the State Level", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-05-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "XBOUND evaluates GUI agents at the level of a screen state, considering multiple instructions and possible target actions within it. The study separates action matching from instruction unification and compares effects of grounding and trajectory data.", "topics": ["evaluation", "grounding"], "updated_at": "2025-09-26", "url": "https://arxiv.org/abs/2505.21279"}
{"aliases": [], "authors": ["Han Xiao", "Guozhi Wang", "Yuxiang Chai", "Zimu Lu", "Weifeng Lin", "Hao He", "Lue Fan", "Liuyang Bian", "Rui Hu", "Liang Liu", "Shuai Ren", "Yafei Wen", "Xiaoxin Chen", "Aojun Zhou", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f19671a527de76b702e68311656e42feea88249716f68911a41ec80be205e4c", "submitted_at": "2025-05-27T17:58:06Z", "updated_at": "2025-05-27T17:58:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.21496v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.21496", "identifiers": {"arxiv": "2505.21496"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21496"}], "name": "UI-Genie: A Self-Improving Approach for Iteratively Boosting MLLM-based Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-Genie jointly improves a mobile GUI agent and a reward model that processes screenshots with interaction history and supplies action- and task-level feedback. Its data pipeline uses rule-based verification, corrupted trajectories, and hard negatives to support reward-guided exploration and iterative training.", "topics": ["data-generation", "reinforcement-learning", "mobile-navigation", "interaction-history"], "updated_at": "2025-05-27", "url": "https://arxiv.org/abs/2505.21496"}
{"aliases": [], "authors": ["Haowei Wang", "Junjie Wang", "Xiaojun Jia", "Rupeng Zhang", "Mingyang Li", "Zhe Liu", "Yang Liu", "Qing Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8aade17e67471c87f259d49aef43b3caf764b6ac14f99e1a4114f4a25dda3d6", "submitted_at": "2025-05-27T17:59:05Z", "updated_at": "2025-05-27T17:59:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.21499v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.21499", "identifiers": {"arxiv": "2505.21499"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21499"}], "name": "AdInject: Real-World Black-Box Attacks on Web Agents via Advertising Delivery", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-05-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies advertising delivery as an indirect-injection channel for web agents under a black-box threat model. AdInject evaluates how static ad content shaped by website context can redirect browser actions.", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2025-05-27", "url": "https://arxiv.org/abs/2505.21499"}
{"aliases": [], "authors": ["Zeyi Liao", "Jaylen Jones", "Linxi Jiang", "Yuting Ning", "Eric Fosler-Lussier", "Yu Su", "Zhiqiang Lin", "Huan Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fe3bdb1357a375de0b1700a6eda7d5987f358bd47fddf3405c27272a3453f2f", "submitted_at": "2025-05-28T03:42:09Z", "updated_at": "2026-03-01T07:50:09Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2505.21936v5"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.21936", "identifiers": {"arxiv": "2505.21936"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21936"}], "name": "RedTeamCUA: Realistic Adversarial Testing of Computer-Use Agents in Hybrid Web-OS Environments", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "RedTeamCUA combines a virtual-machine OS environment with containerized websites for controlled prompt-injection tests spanning both interfaces. Its RTC-Bench contains 864 examples and can initialize agents at injection points to separate attack susceptibility from navigation failures.", "topics": ["safety-security", "evaluation", "sandbox"], "updated_at": "2026-03-01", "url": "https://arxiv.org/abs/2505.21936"}
{"aliases": [], "authors": ["Ziyun Zhang", "Xinyi Liu", "Xiaoyi Zhang", "Jun Wang", "Gang Chen", "Yan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "17c0dac4cbb7c3b7b94496dc225db061a0dc3bda0a34ef40c58ea451524fada7", "submitted_at": "2025-05-28T04:32:05Z", "updated_at": "2025-11-03T08:44:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.21964v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.21964", "identifiers": {"arxiv": "2505.21964"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21964"}], "name": "UI-Evol: Automatic Knowledge Evolving for Computer Use Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-05-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces UI-Evol, which extracts action sequences from agent interactions and critiques external guidance against those traces. Evaluation on OSWorld examines how revised operational knowledge affects execution and behavioral consistency.", "topics": ["memory", "self-reflection", "robustness"], "updated_at": "2025-11-03", "url": "https://arxiv.org/abs/2505.21964"}
{"aliases": [], "authors": ["Jiali Chen", "Xusen Hei", "HongFei Liu", "Yuancheng Wei", "Zikun Deng", "Jiayuan Xie", "Yi Cai", "Li Qing"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3b21d81b786afb70bfcc388082ac6f92fe728bf2383eaf9e61bdf9599f8dea33", "submitted_at": "2025-05-28T12:41:00Z", "updated_at": "2025-05-28T12:41:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.22304v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.22304", "identifiers": {"arxiv": "2505.22304"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.22304"}], "name": "CADReview: Automatically Reviewing CAD Programs with Error Detection and Correction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-28", "url": "https://arxiv.org/abs/2505.22304"}
{"aliases": [], "authors": ["Rong Li", "Shijie Li", "Lingdong Kong", "Xulei Yang", "Junwei Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a4a784e6849a4191603d40ffd5f13cd7900c73063bec7a2aa3f939355ba4b9df", "submitted_at": "2025-05-28T14:53:53Z", "updated_at": "2025-05-28T14:53:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.22429v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.22429", "identifiers": {"arxiv": "2505.22429"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.22429"}], "name": "Zero-Shot 3D Visual Grounding from Vision-Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-28", "url": "https://arxiv.org/abs/2505.22429"}
{"aliases": [], "authors": ["Jialong Wu", "Baixuan Li", "Runnan Fang", "Wenbiao Yin", "Liwen Zhang", "Zhengwei Tao", "Dingchu Zhang", "Zekun Xi", "Gang Fu", "Yong Jiang", "Pengjun Xie", "Fei Huang", "Jingren Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52a13783908f1e85bd50b95b319ada61ed902102f595f808105673a59d4318f5", "submitted_at": "2025-05-28T17:57:07Z", "updated_at": "2025-08-10T06:05:46Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.22648v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2505.22648", "identifiers": {"arxiv": "2505.22648"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.22648"}], "name": "WebDancer: Towards Autonomous Information Seeking Agency", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-10", "url": "https://arxiv.org/abs/2505.22648"}
{"aliases": [], "authors": ["Yuchen Zhuang", "Di Jin", "Jiaao Chen", "Wenqi Shi", "Hanrui Wang", "Chao Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9d260f52763eac186e6f0ff614c818f17602658b6be27cfecc3fea9d80579c9a", "submitted_at": "2025-05-28T23:45:28Z", "updated_at": "2025-06-08T07:57:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.22942v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2505.22942", "identifiers": {"arxiv": "2505.22942"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.22942"}], "name": "WorkForceAgent-R1: Incentivizing Reasoning Capability in LLM-based Web Agents via Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses rule-based rewards for output structure and action correctness to train reasoning for enterprise web navigation.", "topics": ["reinforcement-learning", "workflow"], "updated_at": "2025-06-08", "url": "https://arxiv.org/abs/2505.22942"}
{"aliases": [], "authors": ["Zhejian Yang", "Yongchao Chen", "Xueyang Zhou", "Jiangyue Yan", "Dingjie Song", "Yinuo Liu", "Yuting Li", "Yu Zhang", "Pan Zhou", "Hechang Chen", "Lichao Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "111a0ca1c6ef73a6cea87df28fd43d891a109d0a50b57d1e80463ae761354a90", "submitted_at": "2025-05-29T13:56:49Z", "updated_at": "2025-06-11T03:16:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.23450v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.23450", "identifiers": {"arxiv": "2505.23450"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.23450"}], "name": "Agentic Robot: A Brain-Inspired Framework for Vision-Language-Action Models in Embodied Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-11", "url": "https://arxiv.org/abs/2505.23450"}
{"aliases": [], "authors": ["Kunlun Zhu", "Jiaxun Zhang", "Ziheng Qi", "Nuoxing Shang", "Zijia Liu", "Peixuan Han", "Yue Su", "Haofei Yu", "Jiaxuan You"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b491bf867f7e8b180f247b542668868f6ecabdc3b128b0b28e38d23a54f212c7", "submitted_at": "2025-05-29T15:35:58Z", "updated_at": "2025-05-29T15:35:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.23559v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2505.23559", "identifiers": {"arxiv": "2505.23559"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2505.23559"}], "name": "SafeScientist: Toward Risk-Aware Scientific Discoveries by LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-05-29", "url": "https://arxiv.org/abs/2505.23559"}
{"aliases": [], "authors": ["Linqiang Guo", "Wei Liu", "Yi Wen Heng", "Tse-Hsun", "Chen", "Yang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f0ec81429fb39e56414d3bfe621df6c6adf461e63fc1a1c8886ea2b6850b8ee", "submitted_at": "2025-05-29T16:08:51Z", "updated_at": "2025-11-19T16:48:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.23596v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2505.23596", "identifiers": {"arxiv": "2505.23596"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.23596"}], "name": "Agent-SAMA: State-Aware Mobile Assistant", "notes": "Hugging Face discovery used the label MAPLE for this ID; the record follows the current primary arXiv title, Agent-SAMA. The secondary-title discrepancy is not treated as proof of a historical rename. Experiments not independently reproduced.", "published_at": "2025-05-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Models mobile application execution as a finite-state machine to support planning, verification, and recovery across interface transitions.", "topics": ["planning", "memory", "mobile-navigation"], "updated_at": "2025-11-19", "url": "https://arxiv.org/abs/2505.23596"}
{"aliases": [], "authors": ["Chenyu Yang", "Shiqian Su", "Shi Liu", "Xuan Dong", "Yue Yu", "Weijie Su", "Xuehui Wang", "Zhaoyang Liu", "Jinguo Zhu", "Hao Li", "Wenhai Wang", "Yu Qiao", "Xizhou Zhu", "Jifeng Dai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "799e7b387f7d281f0aa9b182f37b215e3baca2041a0a0f6d414ddf1cec0de003", "submitted_at": "2025-05-29T17:59:51Z", "updated_at": "2025-05-29T17:59:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.23762v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2505.23762", "identifiers": {"arxiv": "2505.23762"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.23762"}], "name": "ZeroGUI: Automating Online GUI Learning at Zero Human Cost", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-05-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines generated tasks, estimated success rewards, and online reinforcement learning to automate GUI-agent interaction training.", "topics": ["reinforcement-learning", "data-generation"], "updated_at": "2025-05-29", "url": "https://arxiv.org/abs/2505.23762"}
{"aliases": [], "authors": ["Brandon Man", "Ghadi Nehme", "Md Ferdous Alam", "Faez Ahmed"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f913b87f07fdad0e4c45264c58c8d9b9d13ca26695306d9e0644eb856f18ced7", "submitted_at": "2025-05-30T17:39:52Z", "updated_at": "2025-11-08T18:27:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.24838v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:34.311949+00:00", "role": "primary", "sha256": "a303d95a811331cd6ee23b0340f1b9ef2dd996835bfbc66656a871ff94eead50", "supports": ["environments", "notes"], "url": "https://arxiv.org/html/2505.24838v2"}], "id": "arxiv:2505.24838", "identifiers": {"arxiv": "2505.24838"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.24838"}], "name": "VideoCAD: A Dataset and Model for Learning Long-Horizon 3D CAD UI Interactions from Video", "notes": "The contribution was reviewed from the retained abstract. Sections 1 and 3 of the exact-version v2 paper identify Onshape as the browser-based interaction environment; the use of pixel-level input does not make the operated surface a desktop application. No experimental results were independently reproduced.", "published_at": "2025-05-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces synthetic CAD interaction videos, a model for learning interface actions from video, and evaluation tasks for spatial and long-horizon reasoning.", "topics": ["video", "data-generation", "science", "long-horizon"], "updated_at": "2025-11-08", "url": "https://arxiv.org/abs/2505.24838"}
{"aliases": [], "authors": ["Yaxin Luo", "Zhaoyi Li", "Jiacheng Liu", "Jiacheng Cui", "Xiaohan Zhao", "Zhiqiang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f2206333ca25b3f21df7ce49adfd8bc56521b9b8ace7a9b686eccf5b67276fa", "submitted_at": "2025-05-30T17:59:55Z", "updated_at": "2025-05-30T17:59:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.24878v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2505.24878", "identifiers": {"arxiv": "2505.24878"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.24878"}], "name": "Open CaptchaWorld: A Comprehensive Web-based Platform for Testing and Benchmarking Multimodal LLM Agents", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-05-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Open CaptchaWorld evaluates multimodal agents on interactive browser-based CAPTCHA puzzles. Its benchmark distinguishes task success from the depth of visual reasoning and interaction needed to complete a puzzle.", "topics": ["evaluation", "web-navigation", "screen-understanding"], "updated_at": "2025-05-30", "url": "https://arxiv.org/abs/2505.24878"}
{"aliases": [], "authors": ["Jingyi Yang", "Shuai Shao", "Dongrui Liu", "Jing Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb94fdba775cb0fb54b312fdff9fa71afd3aefa56b812ff31ceccad33eb5f800", "submitted_at": "2025-05-31T16:04:59Z", "updated_at": "2025-06-20T01:24:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.00618v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.00618", "identifiers": {"arxiv": "2506.00618"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.00618"}], "name": "RiOSWorld: Benchmarking the Risk of Multimodal Computer-Use Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-05-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces RiOSWorld, a computer-use risk benchmark separating user-originated and environmental threats. It evaluates both risky-goal intent and goal completion across web and other computer applications.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-06-20", "url": "https://arxiv.org/abs/2506.00618"}
{"aliases": [], "authors": ["Hanjun Luo", "Shenyu Dai", "Chiming Ni", "Xinfeng Li", "Guibin Zhang", "Kun Wang", "Tongliang Liu", "Hanan Salam"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f62d56a8dd282212ef34756cfac17281908a59478a65e988f083260c29a03a93", "submitted_at": "2025-05-31T17:10:23Z", "updated_at": "2026-01-31T00:21:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.00641v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}], "id": "arxiv:2506.00641", "identifiers": {"arxiv": "2506.00641"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.00641"}], "name": "AgentAuditor: Human-Level Safety and Security Evaluation for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-05-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-31", "url": "https://arxiv.org/abs/2506.00641"}
{"aliases": [], "authors": ["Zhong Zhang", "Yaxi Lu", "Yikun Fu", "Yupeng Huo", "Shenzhi Yang", "Yesai Wu", "Han Si", "Xin Cong", "Haotian Chen", "Yankai Lin", "Jie Xie", "Wei Zhou", "Wang Xu", "Yuanheng Zhang", "Zhou Su", "Zhongwu Zhai", "Xiaoming Liu", "Yudong Mei", "Jianming Xu", "Hongyan Tian", "Chongyi Wang", "Chi Chen", "Yuan Yao", "Zhiyuan Liu", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c0b2a065898ba4f73fcb5e5b2e92d723b0b55ac6caa46d5bf02d44eb7a2cdd9", "submitted_at": "2025-06-02T07:30:29Z", "updated_at": "2025-06-17T02:57:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.01391v2"}, "publication": {"authors": ["Zhong Zhang", "Yaxi Lu", "Yikun Fu", "Yupeng Huo", "Shenzhi Yang", "Yesai Wu", "Han Si (司函)", "Xin Cong", "Haotian Chen", "Yankai Lin (林衍凯)", "Jie Xie", "Wei Zhou", "Wang Xu", "Yuanheng Zhang", "Zhou Su", "Zhongwu Zhai", "Xiaoming Liu", "Yudong Mei", "Jianming Xu", "Hongyan Tian", "Chongyi Wang", "Chi Chen", "Yuan Yao", "Zhiyuan Liu", "Maosong Sun (孙茂松)"], "date_basis": "Explicit publisher citation_publication_date: 2025/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 11], "doi": "10.18653/v1/2025.emnlp-demos.12", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.emnlp-demos.12/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "155-180", "pdf_url": "https://aclanthology.org/2025.emnlp-demos.12.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AgentCPM-GUI: Building Mobile-Use Agents with Reinforcement Fine-Tuning", "type": "conference-paper", "url": "https://aclanthology.org/2025.emnlp-demos.12/", "venue": "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: System Demonstrations", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:59Z", "role": "primary", "sha256": "5de5a828e06a63c1627b8ad9f9a46c1643d8b4b7d8ed6156fd9790b6bd083155", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.emnlp-demos.12/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2506.01391", "identifiers": {"arxiv": "2506.01391"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.01391"}, {"rel": "publication", "url": "https://aclanthology.org/2025.emnlp-demos.12/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.emnlp-demos.12.pdf"}], "name": "AgentCPM-GUI: Building Mobile-Use Agents with Reinforcement Fine-Tuning", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-06-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines grounding pretraining, bilingual interaction trajectories, and reinforcement fine-tuning for mobile GUI actions.", "topics": ["reinforcement-learning", "mobile-navigation"], "updated_at": "2025-06-17", "url": "https://arxiv.org/abs/2506.01391"}
{"aliases": [], "authors": ["Hongyang Yang", "Likun Lin", "Yang She", "Xinyu Liao", "Jiaoyang Wang", "Runjia Zhang", "Yuquan Mo", "Christina Dan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8da91610f3e41dfd058e4db3cd127d9ed79507799fec286bd08314c46c430d19", "submitted_at": "2025-06-02T08:22:28Z", "updated_at": "2025-06-02T08:22:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.01423v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.01423", "identifiers": {"arxiv": "2506.01423"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.01423"}], "name": "FinRobot: Generative Business Process AI Agents for Enterprise Resource Planning in Finance", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-02", "url": "https://arxiv.org/abs/2506.01423"}
{"aliases": [], "authors": ["Xiao Yang", "Jiawei Chen", "Jun Luo", "Zhengwei Fang", "Yinpeng Dong", "Hang Su", "Jun Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ec2467703bc5d0f97ef8d207289ed56f6d12b811688d20b7833eac06038757b", "submitted_at": "2025-06-02T12:56:27Z", "updated_at": "2025-06-02T12:56:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.01616v1"}}, "environments": ["web", "android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:55Z", "role": "primary", "sha256": "d69da19edaed185ad3071a97db3bb3fcfbdb14c0b514b7d27794b39fbe8e2541", "supports": ["environments", "notes"], "url": "https://raw.githubusercontent.com/thu-ml/MLA-Trust/HEAD/README.md"}], "id": "arxiv:2506.01616", "identifiers": {"arxiv": "2506.01616"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.01616"}], "name": "MLA-Trust: Benchmarking Trustworthiness of Multimodal LLM Agents in GUI Environments", "notes": "The contribution was reviewed from the retained paper abstract; Android setup is supported separately by the official project README. No benchmark results were independently reproduced.", "published_at": "2025-06-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Benchmarks truthfulness, controllability, safety, and privacy of multimodal LLM agents across interactive web and mobile GUI tasks.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2025-06-02", "url": "https://arxiv.org/abs/2506.01616"}
{"aliases": [], "authors": ["Peter Belcak", "Greg Heinrich", "Shizhe Diao", "Yonggan Fu", "Xin Dong", "Saurav Muralidharan", "Yingyan Celine Lin", "Pavlo Molchanov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a83ee6f8b372e330cf421a6463bd9deaa348a6bfa37e80e0892e4ed9866469e0", "submitted_at": "2025-06-02T18:35:16Z", "updated_at": "2026-09-22T02:53:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.02153v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2506.02153", "identifiers": {"arxiv": "2506.02153"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.02153"}], "name": "Small Language Models are the Future of Agentic AI", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-22", "url": "https://arxiv.org/abs/2506.02153"}
{"aliases": [], "authors": ["Tri Cao", "Bennett Lim", "Yue Liu", "Yuan Sui", "Yuexin Li", "Shumin Deng", "Lin Lu", "Nay Oo", "Shuicheng Yan", "Bryan Hooi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3cd988217d1e3371bec827b472abf46c205ac407b5b60b7b581d7105f86e8bc5", "submitted_at": "2025-06-03T05:21:50Z", "updated_at": "2026-03-01T16:12:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.02456v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.02456", "identifiers": {"arxiv": "2506.02456"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.02456"}], "name": "VPI-Bench: Visual Prompt Injection Attacks for Computer-Use Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces VPI-Bench, interactive web-platform variants containing visually embedded malicious instructions. It evaluates computer-use and browser-use agents under these threats and examines system-prompt defenses.", "topics": ["safety-security", "evaluation"], "updated_at": "2026-03-01", "url": "https://arxiv.org/abs/2506.02456"}
{"aliases": [], "authors": ["Thong Q. Nguyen", "Shubhang Desai", "Raja Hasnain Anwar", "Firoz Shaik", "Vishwas Suryanarayanan", "Vishal Chowdhary"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "401ca18cadbf38045fa20985354988571b69e2280b1a090aff0e128cdf3c70e0", "submitted_at": "2025-06-03T07:25:49Z", "updated_at": "2025-08-07T21:48:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.02539v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.02539", "identifiers": {"arxiv": "2506.02539"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.02539"}], "name": "VerificAgent: Domain-Specific Memory Verification for Scalable Oversight of Aligned Computer-Use Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces VerificAgent, which grows domain-specific memories from task trajectories and subjects them to human fact-checking before deployment. Its OSWorld evaluation treats persistent memory as an auditable constraint on later agent behavior.", "topics": ["memory", "human-in-the-loop", "safety-security"], "updated_at": "2025-08-07", "url": "https://arxiv.org/abs/2506.02539"}
{"aliases": [], "authors": ["Yougang Lyu", "Xiaoyu Zhang", "Lingyong Yan", "Maarten de Rijke", "Zhaochun Ren", "Xiuying Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bebdf1c8f34582408ef0050245f6b1e454d1bd2e75903a4e0817a4f3c00fb4db", "submitted_at": "2025-06-03T13:08:17Z", "updated_at": "2025-06-03T13:08:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.02839v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2506.02839", "identifiers": {"arxiv": "2506.02839"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.02839"}], "name": "DeepShop: A Benchmark for Deep Research Shopping Agents", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "DeepShop evaluates shopping agents with evolving query complexity across product attributes, filters and sorting preferences. Its evaluation combines checks of these individual requirements with overall task success.", "topics": ["evaluation", "web-navigation"], "updated_at": "2025-06-03", "url": "https://arxiv.org/abs/2506.02839"}
{"aliases": [], "authors": ["Mathieu Andreux", "Breno Baldas Skuk", "Hamza Benchekroun", "Emilien Biré", "Antoine Bonnet", "Riaz Bordie", "Nathan Bout", "Matthias Brunel", "Pierre-Louis Cedoz", "Antoine Chassang", "Mickaël Chen", "Alexandra D. Constantinou", "Antoine d'Andigné", "Hubert de La Jonquière", "Aurélien Delfosse", "Ludovic Denoyer", "Alexis Deprez", "Augustin Derupti", "Michael Eickenberg", "Mathïs Federico", "Charles Kantor", "Xavier Koegler", "Yann Labbé", "Matthew C. H. Lee", "Erwan Le Jumeau de Kergaradec", "Amir Mahla", "Avshalom Manevich", "Adrien Maret", "Charles Masson", "Rafaël Maurin", "Arturo Mena", "Philippe Modard", "Axel Moyal", "Axel Nguyen Kerbel", "Julien Revelle", "Mats L. Richter", "María Santos", "Laurent Sifre", "Maxime Theillard", "Marc Thibault", "Louis Thiry", "Léo Tronchon", "Nicolas Usunier", "Tony Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89c2d658e03204b56f6d273da9572820f9a23ca359a441ea87baa451a93c5b58", "submitted_at": "2025-06-03T13:29:03Z", "updated_at": "2025-06-11T09:33:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.02865v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.02865", "identifiers": {"arxiv": "2506.02865"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.02865"}], "name": "Surfer-H Meets Holo1: Cost-Efficient Web Agent Powered by Open Weights", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes Surfer-H, a web agent paired with the Holo1 vision-language model family trained on web content, synthetic examples, and agent-generated data. The work introduces WebClick for web-interface localization and evaluates browser task execution on WebVoyager.", "topics": ["web-navigation", "grounding", "data-generation", "evaluation"], "updated_at": "2025-06-11", "url": "https://arxiv.org/abs/2506.02865"}
{"aliases": [], "authors": ["Man Luo", "David Cobbley", "Xin Su", "Shachar Rosenman", "Vasudev Lal", "Shao-Yen Tseng", "Phillip Howard"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "068eae2cab9899967850f21616a701f3eeb258b9654c83d34f68be266e48854e", "submitted_at": "2025-06-03T17:27:04Z", "updated_at": "2025-06-03T17:27:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.03095v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.03095", "identifiers": {"arxiv": "2506.03095"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.03095"}], "name": "DPO Learning with LLMs-Judge Signal for Computer Use Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses an LLM judge to evaluate and filter synthetic GUI interaction trajectories for training a compact vision-language agent. OSWorld experiments assess the policy designed for local execution.", "topics": ["data-generation", "evaluation", "efficiency"], "updated_at": "2025-06-03", "url": "https://arxiv.org/abs/2506.03095"}
{"aliases": [], "authors": ["Qianhui Wu", "Kanzhi Cheng", "Rui Yang", "Chaoyun Zhang", "Jianwei Yang", "Huiqiang Jiang", "Jian Mu", "Baolin Peng", "Bo Qiao", "Reuben Tan", "Si Qin", "Lars Liden", "Qingwei Lin", "Huan Zhang", "Tong Zhang", "Jianbing Zhang", "Dongmei Zhang", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87b27d6caaaa3802629fdbf1e7bbe5557ed567f840cee88e45ba7e1fe12c548d", "submitted_at": "2025-06-03T17:59:08Z", "updated_at": "2025-06-03T17:59:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.03143v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.03143", "identifiers": {"arxiv": "2506.03143"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.03143"}], "name": "GUI-Actor: Coordinate-Free Visual Grounding for GUI Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses an attention-based action head and a verifier to select relevant screen regions without generating coordinates as text.", "topics": ["grounding"], "updated_at": "2025-06-03", "url": "https://arxiv.org/abs/2506.03143"}
{"aliases": [], "authors": ["Yajie Zhou", "Jiajun Ruan", "Eric S. Wang", "Sadjad Fouladi", "Francis Y. Yan", "Kevin Hsieh", "Zaoxing Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cd16221e7bdfb06a4acd96a4f2348597d789483b5f9b50d24ff660334b00c570", "submitted_at": "2025-06-03T14:04:22Z", "updated_at": "2026-03-13T19:41:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.03231v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.03231", "identifiers": {"arxiv": "2506.03231"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.03231"}], "name": "NetArena: Dynamic Benchmarks for AI Agents in Network Automation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-13", "url": "https://arxiv.org/abs/2506.03231"}
{"aliases": [], "authors": ["Apurva Gandhi", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bccee69e390b5170ff264e689c209d13317e7d3be9b37e6d3fbed0e0d089943", "submitted_at": "2025-06-04T03:27:56Z", "updated_at": "2026-03-02T22:31:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.03533v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.03533", "identifiers": {"arxiv": "2506.03533"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.03533"}], "name": "Go-Browse: Training Web Agents with Structured Exploration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Go-Browse frames web-environment exploration as graph search so that information can be reused across data-collection episodes. It collects 10,000 successful trajectories and 40,000 interaction steps on WebArena and uses them to fine-tune a web agent.", "topics": ["data-generation", "web-navigation"], "updated_at": "2026-03-02", "url": "https://arxiv.org/abs/2506.03533"}
{"aliases": [], "authors": ["Xiaomi LLM-Core Team", ":", "Zihao Yue", "Zhenru Lin", "Yifan Song", "Weikun Wang", "Shuhuai Ren", "Shuhao Gu", "Shicheng Li", "Peidian Li", "Liang Zhao", "Lei Li", "Kainan Bao", "Hao Tian", "Hailin Zhang", "Gang Wang", "Dawei Zhu", "Cici", "Chenhong He", "Bowen Ye", "Bowen Shen", "Zihan Zhang", "Zihan Jiang", "Zhixian Zheng", "Zhichao Song", "Zhenbo Luo", "Yue Yu", "Yudong Wang", "Yuanyuan Tian", "Yu Tu", "Yihan Yan", "Yi Huang", "Xu Wang", "Xinzhe Xu", "Xingchen Song", "Xing Zhang", "Xing Yong", "Xin Zhang", "Xiangwei Deng", "Wenyu Yang", "Wenhan Ma", "Weiwei Lv", "Weiji Zhuang", "Wei Liu", "Sirui Deng", "Shuo Liu", "Shimao Chen", "Shihua Yu", "Shaohui Liu", "Shande Wang", "Rui Ma", "Qiantong Wang", "Peng Wang", "Nuo Chen", "Menghang Zhu", "Kangyang Zhou", "Kang Zhou", "Kai Fang", "Jun Shi", "Jinhao Dong", "Jiebao Xiao", "Jiaming Xu", "Huaqiu Liu", "Hongshen Xu", "Heng Qu", "Haochen Zhao", "Hanglong Lv", "Guoan Wang", "Duo Zhang", "Dong Zhang", "Di Zhang", "Chong Ma", "Chang Liu", "Can Cai", "Bingquan Xia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc758c8428d7265d34ff261790d97c979daf93599f76d3bfa7735603280ef742", "submitted_at": "2025-06-04T04:32:54Z", "updated_at": "2025-06-04T04:32:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.03569v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2506.03569", "identifiers": {"arxiv": "2506.03569"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.03569"}], "name": "MiMo-VL Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-04", "url": "https://arxiv.org/abs/2506.03569"}
{"aliases": [], "authors": ["Pei Yang", "Hai Ci", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a4c980694819dc5eb4ff322e6a703cccfe0b1e45efb60f22b120592d03edf36", "submitted_at": "2025-06-04T16:26:56Z", "updated_at": "2025-10-18T12:11:31Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2506.04135v4"}}, "environments": ["macos", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.04135", "identifiers": {"arxiv": "2506.04135"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.04135"}], "name": "macOSWorld: A Multilingual Interactive Benchmark for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "macOSWorld provides 202 interactive tasks across 30 applications, with instructions and operating-system interfaces in five languages. It evaluates GUI agents on macOS workflows and includes a separate subset for deception-attack testing.", "topics": ["evaluation", "workflow", "safety-security"], "updated_at": "2025-10-18", "url": "https://arxiv.org/abs/2506.04135"}
{"aliases": [], "authors": ["Jiin Kim", "Byeongjun Shin", "Jinha Chung", "Minsoo Rhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2cbc0999dd570abe5507bb75bd64bacd4e193e78473003d98d5b403555acf950", "submitted_at": "2025-06-04T14:37:54Z", "updated_at": "2026-01-07T00:48:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.04301v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.04301", "identifiers": {"arxiv": "2506.04301"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.04301"}], "name": "The Cost of Dynamic Reasoning: Demystifying AI Agents and Test-Time Scaling from an AI Infrastructure Perspective", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-07", "url": "https://arxiv.org/abs/2506.04301"}
{"aliases": [], "authors": ["Yuyang Wanyan", "Xi Zhang", "Haiyang Xu", "Haowei Liu", "Junyang Wang", "Jiabo Ye", "Yutong Kou", "Ming Yan", "Fei Huang", "Xiaoshan Yang", "Weiming Dong", "Changsheng Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "42be944224a9b7d2c8e7042968ab31d373a0bafa304c0e669728cb0f10cc813d", "submitted_at": "2025-06-05T04:12:36Z", "updated_at": "2025-11-17T07:48:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.04614v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2506.04614", "identifiers": {"arxiv": "2506.04614"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.04614"}], "name": "Look Before You Leap: A GUI-Critic-R1 Model for Pre-Operative Error Diagnosis in GUI Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses a pre-execution critic and reinforcement training to evaluate proposed GUI actions and provide corrective suggestions before execution.", "topics": ["planning", "robustness"], "updated_at": "2025-11-17", "url": "https://arxiv.org/abs/2506.04614"}
{"aliases": [], "authors": ["Shengcheng Yu", "Yuchen Ling", "Chunrong Fang", "Quan Zhou", "Yi Zhao", "Chunyang Chen", "Shaomin Zhu", "Zhenyu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bad447e8bca06f720a9476492de06978c73a356a38b00671203f8af0b81f1952", "submitted_at": "2025-06-05T14:27:40Z", "updated_at": "2026-05-03T08:25:53Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2506.05079v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2506.05079", "identifiers": {"arxiv": "2506.05079"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.05079"}], "name": "Scenario-Guided LLM-based Mobile App GUI Testing", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-06-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "ScenGen organizes mobile GUI testing around business scenarios rather than unrestricted exploration. Five collaborating agents observe the interface, choose and execute actions, check scenario completion, and record operations and runtime bugs.", "topics": ["automation", "mobile-navigation", "planning"], "updated_at": "2026-05-03", "url": "https://arxiv.org/abs/2506.05079"}
{"aliases": [], "authors": ["Ghazi Shazan Ahmad", "Ahmed Heakl", "Hanan Gani", "Abdelrahman Shaker", "Zhiqiang Shen", "Fahad Shahbaz Khan", "Salman Khan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fff25339852c6a0c513b8ccd02dbe447c5bae69eafdec284ce6a9cbfc4d9d943", "submitted_at": "2025-06-05T17:59:29Z", "updated_at": "2025-07-05T11:38:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.05336v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.05336", "identifiers": {"arxiv": "2506.05336"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.05336"}], "name": "VideoMolmo: Spatio-Temporal Grounding Meets Pointing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-05", "url": "https://arxiv.org/abs/2506.05336"}
{"aliases": [], "authors": ["Zihan Deng", "Changyu Du", "Stavros Nousias", "André Borrmann"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bf4f91de32fdf3bd9b81ba64ace02e4a7d02ff094794f30dfd560389f334cea", "submitted_at": "2025-06-08T16:45:31Z", "updated_at": "2025-06-30T08:31:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.07217v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.07217", "identifiers": {"arxiv": "2506.07217"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.07217"}], "name": "BIMgent: Towards Autonomous Building Modeling via Computer-use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "BIMgent plans software-specific workflows and executes GUI actions for building-information modeling. It evaluates multimodal building design and reconstruction tasks in professional authoring software.", "topics": ["industry", "planning", "workflow", "evaluation"], "updated_at": "2025-06-30", "url": "https://arxiv.org/abs/2506.07217"}
{"aliases": [], "authors": ["Ruofan Liu", "Xiwen Teoh", "Yun Lin", "Guanjie Chen", "Ruofei Ren", "Denys Poshyvanyk", "Jin Song Dong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bea3e76b3124ebbcd8f7817b3e32d667d8aaae6e31f1b75ac3fc8c39e7637c6f", "submitted_at": "2025-06-09T03:09:48Z", "updated_at": "2025-06-09T03:09:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.07385v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:30Z", "role": "primary", "sha256": "23030fd266519162c51c7f4a87bc38e75bb171ef7fa95e30c2d7afd820a0bc3f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2506.07385v1"}], "id": "arxiv:2506.07385", "identifiers": {"arxiv": "2506.07385"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.07385"}], "name": "GUIPilot: A Consistency-based Mobile GUI Testing Approach for Detecting Application-specific Bugs", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUIPilot, which compares mobile design mock-ups with implemented screens through widget alignment and tests expected transitions by executing inferred UI actions. It evaluates both visual inconsistencies and workflow-specific bugs.", "topics": ["evaluation", "mobile-navigation", "robustness"], "updated_at": "2025-06-09", "url": "https://arxiv.org/abs/2506.07385"}
{"aliases": [], "authors": ["Guibin Zhang", "Muxin Fu", "Guancheng Wan", "Miao Yu", "Kun Wang", "Shuicheng Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ca25df31fbf4ae92f6874dc5c0b3860e87815459bbf1e4a4417ddfd1a14cf84", "submitted_at": "2025-06-09T03:43:46Z", "updated_at": "2025-06-16T08:45:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.07398v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.07398", "identifiers": {"arxiv": "2506.07398"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.07398"}], "name": "G-Memory: Tracing Hierarchical Memory for Multi-Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-16", "url": "https://arxiv.org/abs/2506.07398"}
{"aliases": [], "authors": ["Yunhe Yan", "Shihe Wang", "Jiajun Du", "Yexuan Yang", "Yuxuan Shan", "Qichen Qiu", "Xianqing Jia", "Xinge Wang", "Xin Yuan", "Xu Han", "Mao Qin", "Yinxiao Chen", "Chen Peng", "Shangguang Wang", "Mengwei Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2334e017fe49b39dc5d460963988f05d11051f2101bebe23be76005809fafd45", "submitted_at": "2025-06-09T11:50:33Z", "updated_at": "2025-06-09T11:50:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.07672v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:12:09Z", "role": "primary", "sha256": "f64464b1516369a87f0fbe28828e32ff61687dae4ce063aedddd8bd45a7d40eb", "supports": ["scope", "summary", "environments"], "url": "https://arxiv.org/html/2506.07672"}], "id": "arxiv:2506.07672", "identifiers": {"arxiv": "2506.07672"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.07672"}], "name": "MCPWorld: A Unified Benchmarking Testbed for API, GUI, and Hybrid Computer Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The platform or GUI evaluation setting was additionally checked in the primary full text.", "published_at": "2025-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "MCPWorld evaluates API, GUI, and hybrid computer-use agents with 201 tasks in applications whose source can be instrumented and extended with MCP support. It verifies completion from application behavior and packages the testbed in containers with GPU support.", "topics": ["evaluation", "tool-integration", "sandbox"], "updated_at": "2025-06-09", "url": "https://arxiv.org/abs/2506.07672"}
{"aliases": [], "authors": ["Junhong Shen", "Hao Bai", "Lunjun Zhang", "Yifei Zhou", "Amrith Setlur", "Shengbang Tong", "Diego Caples", "Nan Jiang", "Tong Zhang", "Ameet Talwalkar", "Aviral Kumar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4b6c8b297e88dae689ba2a48724a441698cd5309a3bbe1198f38343e88aceb1", "submitted_at": "2025-06-09T17:50:02Z", "updated_at": "2025-06-10T12:50:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.07976v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2506.07976", "identifiers": {"arxiv": "2506.07976"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.07976"}], "name": "Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interaction", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Test-Time Interaction scales the interaction horizon so web agents can explore, backtrack and revise plans within a rollout. Its online reinforcement-learning curriculum adjusts rollout lengths and is evaluated on WebVoyager and WebArena.", "topics": ["reinforcement-learning", "web-navigation", "planning", "long-horizon"], "updated_at": "2025-06-10", "url": "https://arxiv.org/abs/2506.07976"}
{"aliases": [], "authors": ["Penghao Wu", "Shengnan Ma", "Bo Wang", "Jiaheng Yu", "Lewei Lu", "Ziwei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce3b7d2c25b5a99f110f00be7389d4230d625ac1f7f33d5ab81f3d6f441f3610", "submitted_at": "2025-06-09T17:59:57Z", "updated_at": "2025-06-09T17:59:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.08012v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2506.08012", "identifiers": {"arxiv": "2506.08012"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.08012"}], "name": "GUI-Reflection: Empowering Multimodal GUI Models with Self-Reflection Behavior", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Reflection synthesizes reflection and error-correction examples from successful trajectories and adds dedicated learning stages for these behaviors. It introduces a reflection task suite, a mobile environment for online data collection, and iterative reflection tuning.", "topics": ["robustness", "data-generation", "evaluation", "mobile-navigation"], "updated_at": "2025-06-09", "url": "https://arxiv.org/abs/2506.08012"}
{"aliases": [], "authors": ["Zefang Liu", "Yinzhu Quan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43aca3b0b8a1a0d5729a4e90dc1516bbb68e3611de3220876cdb29afc81d847f", "submitted_at": "2025-06-09T18:39:48Z", "updated_at": "2026-05-11T08:25:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.08136v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:34Z", "role": "primary", "sha256": "2f1d53911fe997bbba01c9d8a2fa7002f18837492448e15056fd7a7a6c78f97d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2506.08136v3"}], "id": "arxiv:2506.08136", "identifiers": {"arxiv": "2506.08136"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.08136"}], "name": "EconWebArena: Benchmarking Autonomous Agents on Economic Tasks in Realistic Web Environments", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces EconWebArena, a curated benchmark of economic-data tasks requiring live website navigation, multimodal interpretation, and precise extraction. It evaluates visual grounding, planning, and interaction design within multistep workflows.", "topics": ["evaluation", "web-navigation", "planning"], "updated_at": "2026-05-11", "url": "https://arxiv.org/abs/2506.08136"}
{"aliases": [], "authors": ["Luca Beurer-Kellner", "Beat Buesser", "Ana-Maria Creţu", "Edoardo Debenedetti", "Daniel Dobos", "Daniel Fabian", "Marc Fischer", "David Froelicher", "Kathrin Grosse", "Daniel Naeff", "Ezinwanne Ozoani", "Andrew Paverd", "Florian Tramèr", "Václav Volhejn"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "305a125528f51f8c0307793e953cfe1252b2fbfc7be12a63cbc040db1bd80d23", "submitted_at": "2025-06-10T14:23:55Z", "updated_at": "2025-06-27T11:13:27Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.08837v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.08837", "identifiers": {"arxiv": "2506.08837"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.08837"}], "name": "Design Patterns for Securing LLM Agents against Prompt Injections", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-27", "url": "https://arxiv.org/abs/2506.08837"}
{"aliases": [], "authors": ["Yuan Guo", "Tingjia Miao", "Zheng Wu", "Pengzhou Cheng", "Ming Zhou", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8dbc84aa582b94775d490cf84e17eff05be84af7dbcc2de7de61b886fa3c9178", "submitted_at": "2025-06-10T16:45:29Z", "updated_at": "2025-06-10T16:45:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.08972v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "id": "arxiv:2506.08972", "identifiers": {"arxiv": "2506.08972"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.08972"}], "name": "Atomic-to-Compositional Generalization for Mobile Agents with A New Benchmark and Scheduling System", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-NEXUS evaluates compositional mobile tasks in local utility apps and online Chinese and English service apps. AGENT-NEXUS decomposes those long tasks into self-contained atomic subtasks to study generalization beyond isolated screens or short action chains.", "topics": ["evaluation", "long-horizon", "planning", "mobile-navigation"], "updated_at": "2025-06-10", "url": "https://arxiv.org/abs/2506.08972"}
{"aliases": [], "authors": ["Jiaqi Tang", "Yu Xia", "Yi-Feng Wu", "Yuwei Hu", "Yuhui Chen", "Qing-Guo Chen", "Xiaogang Xu", "Xiangyu Wu", "Hao Lu", "Yanqing Ma", "Shiyin Lu", "Qifeng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b00deba927bda03c079621f70decdeb946ec51a6982531a2086c43033ac9925c", "submitted_at": "2025-06-11T03:43:30Z", "updated_at": "2026-04-21T11:50:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.09373v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.09373", "identifiers": {"arxiv": "2506.09373"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.09373"}], "name": "LPO: Towards Accurate GUI Agent Interaction via Location Preference Optimization", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Location Preference Optimization uses an information-entropy signal to select interaction regions and a distance-based reward to assess click locations. It trains GUI localization through preference optimization and evaluates both offline grounding and online interaction.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2026-04-21", "url": "https://arxiv.org/abs/2506.09373"}
{"aliases": [], "authors": ["Yuquan Xie", "Zaijing Li", "Rui Shao", "Gongwei Chen", "Kaiwen Zhou", "Yinchuan Li", "Dongmei Jiang", "Liqiang Nie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b353606ce29a5bc60d263e86827a825af46883e25da5bd290b48affa9dd3838", "submitted_at": "2025-06-12T06:21:19Z", "updated_at": "2025-06-12T06:21:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.10387v1"}}, "environments": ["android", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.10387", "identifiers": {"arxiv": "2506.10387"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.10387"}], "name": "Mirage-1: Augmenting and Updating GUI Agent with Hierarchical Multimodal Skills", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Mirage-1 organizes interaction trajectories into execution, core, and meta-skills, then uses them to guide Monte Carlo tree search during online GUI tasks. It introduces AndroidLH for long-horizon evaluation and tests the approach on Android and web-navigation benchmarks.", "topics": ["planning", "memory", "long-horizon", "mobile-navigation", "web-navigation"], "updated_at": "2025-06-12", "url": "https://arxiv.org/abs/2506.10387"}
{"aliases": [], "authors": ["Xing Han Lù", "Gaurav Kamath", "Marius Mosbach", "Siva Reddy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c73f4a7361e714b57de5a49631caa8b342850c79a1a6fa5354ec01b9f8b648ef", "submitted_at": "2025-06-12T17:53:58Z", "updated_at": "2025-06-12T17:53:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.10953v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.10953", "identifiers": {"arxiv": "2506.10953"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.10953"}], "name": "Build the web for agents, not agents for the web", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-12", "url": "https://arxiv.org/abs/2506.10953"}
{"aliases": [], "authors": ["Wenkang Han", "Zhixiong Zeng", "Jing Huang", "Shu Jiang", "Liming Zheng", "Longrong Yang", "Haibo Qiu", "Chang Yao", "Jingyuan Chen", "Lin Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3cce7c2e1623ec841e8afed03c2e8e0c6f969f3dd05d53ddc4ef11e74310c21c", "submitted_at": "2025-06-10T12:16:27Z", "updated_at": "2025-11-26T07:48:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.11127v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.11127", "identifiers": {"arxiv": "2506.11127"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.11127"}], "name": "UITron-Speech: Towards Automated GUI Agents Based on Speech Instructions", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-06-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces UITron-Speech, which predicts GUI actions from spoken instructions and screenshots. It combines synthetic speech data and mixed-modality training with a training-free grounding refinement step.", "topics": ["grounding", "action-prediction", "data-generation"], "updated_at": "2025-11-26", "url": "https://arxiv.org/abs/2506.11127"}
{"aliases": [], "authors": ["Jeff Da", "Clinton Wang", "Xiang Deng", "Yuntao Ma", "Nikhil Barhate", "Sean Hendryx"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ea6180c506bf9266a4bdc65a72fcc827ea30932658391e6ab07581e066302d5", "submitted_at": "2025-06-13T02:46:53Z", "updated_at": "2025-06-20T23:32:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.11425v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.11425", "identifiers": {"arxiv": "2506.11425"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.11425"}], "name": "Agent-RLVR: Training Software Engineering Agents via Guidance and Environment Rewards", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-20", "url": "https://arxiv.org/abs/2506.11425"}
{"aliases": [], "authors": ["Mingxuan Du", "Benfeng Xu", "Chiwei Zhu", "Xiaorui Wang", "Zhendong Mao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "277876a7927da3d89ac083410e338453a83a04c47b30912519e1cd70ccf63d45", "submitted_at": "2025-06-13T13:17:32Z", "updated_at": "2025-06-13T13:17:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.11763v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.11763", "identifiers": {"arxiv": "2506.11763"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.11763"}], "name": "DeepResearch Bench: A Comprehensive Benchmark for Deep Research Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-13", "url": "https://arxiv.org/abs/2506.11763"}
{"aliases": [], "authors": ["Xuan Wang", "Siyuan Liang", "Zhe Liu", "Yi Yu", "Aishan Liu", "Yuliang Lu", "Xitong Gao", "Ee-Chien Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "390304ce515886fa15fb8cd74d8b2df161a75b9adc0f4d2fad1f65976c6a6ce7", "submitted_at": "2025-06-16T08:09:32Z", "updated_at": "2025-09-05T14:19:03Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2506.13205v6"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "id": "arxiv:2506.13205", "identifiers": {"arxiv": "2506.13205"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.13205"}], "name": "Poison Once, Control Anywhere: Clean-Text Visual Backdoors in VLM-based Mobile Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Demonstrates clean-text visual backdoor attacks (VIBMA) that compromise fine-tuned VLM mobile agents through visual triggers while keeping text instructions intact.", "topics": ["safety-security", "mobile-navigation", "robustness"], "updated_at": "2025-09-05", "url": "https://arxiv.org/abs/2506.13205"}
{"aliases": [], "authors": ["Mohammed Mehedi Hasan", "Hao Li", "Emad Fallahzadeh", "Gopi Krishnan Rajbahadur", "Bram Adams", "Ahmed E. Hassan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a11305b4a7e876e1797c93e369c8157a7e9759604e1383839c4f155dc0e941b", "submitted_at": "2025-06-16T14:26:37Z", "updated_at": "2026-04-13T03:45:24Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2506.13538v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.13538", "identifiers": {"arxiv": "2506.13538"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.13538"}], "name": "Model Context Protocol (MCP) at First Glance: Studying the Security and Maintainability of MCP Servers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-13", "url": "https://arxiv.org/abs/2506.13538"}
{"aliases": [], "authors": ["Jingxu Xie", "Dylan Xu", "Xuandong Zhao", "Dawn Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae36b271f9b4799093831cabee772845edfaeca97107414e424e170ccf89904d", "submitted_at": "2025-06-17T05:46:52Z", "updated_at": "2026-03-02T00:51:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.14205v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:13+00:00", "role": "primary", "sha256": "42e8b8dead7caea3f8b21e93a075e8c895dfc30a25e1470bf97a90940bbf0106", "supports": ["scope", "environments"], "url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/README.md"}], "id": "arxiv:2506.14205", "identifiers": {"arxiv": "2506.14205"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.14205"}], "name": "AgentSynth: Scalable Task Generation for Generalist Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-06-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "AgentSynth composes easily generated subtasks into longer computer-use tasks and trajectories with adjustable difficulty. Its evaluation uses the number of subtasks to probe how agent success changes as task complexity grows.", "topics": ["data-generation", "evaluation", "long-horizon"], "updated_at": "2026-03-02", "url": "https://arxiv.org/abs/2506.14205"}
{"aliases": [], "authors": ["Jingqi Yang", "Zhilong Song", "Jiawei Chen", "Mingli Song", "Sheng Zhou", "linjun sun", "Xiaogang Ouyang", "Chun Chen", "Can Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f9e64118f03c2a9f3f55e35f79c38835f1eb823265c83c933fdceca2610b9904", "submitted_at": "2025-06-17T12:50:35Z", "updated_at": "2025-06-17T12:50:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.14477v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.14477", "identifiers": {"arxiv": "2506.14477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.14477"}], "name": "GUI-Robust: A Comprehensive Dataset for Testing GUI Agent Robustness in Real-World Anomalies", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds GUI task data with real-world interaction anomalies and evaluates how agent behavior changes under those conditions.", "topics": ["robustness", "evaluation", "data-generation"], "updated_at": "2025-06-17", "url": "https://arxiv.org/abs/2506.14477"}
{"aliases": [], "authors": ["Leonhard Applis", "Yuntong Zhang", "Shanchao Liang", "Nan Jiang", "Lin Tan", "Abhik Roychoudhury"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "25505a98668fe12a13b4868922229fcbf4634ad0104ba7023db0efcf08267c58", "submitted_at": "2025-06-17T16:19:13Z", "updated_at": "2025-12-08T09:06:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.14683v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.14683", "identifiers": {"arxiv": "2506.14683"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.14683"}], "name": "Unified Software Engineering Agent as AI Software Engineer", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-08", "url": "https://arxiv.org/abs/2506.14683"}
{"aliases": [], "authors": ["Thomas Kuntz", "Agatha Duzan", "Hao Zhao", "Francesco Croce", "Zico Kolter", "Nicolas Flammarion", "Maksym Andriushchenko"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f551b861012fc4eddbf010a619bc65e77b524b05958761b738a88bcae377e9a9", "submitted_at": "2025-06-17T17:59:31Z", "updated_at": "2025-10-29T10:34:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.14866v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.14866", "identifiers": {"arxiv": "2506.14866"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.14866"}], "name": "OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-06-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Extends OSWorld with tasks covering user misuse, prompt injection, and model misbehavior, with assessment of both task success and safety.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-10-29", "url": "https://arxiv.org/abs/2506.14866"}
{"aliases": [], "authors": ["Xingjian Tao", "Yiwei Wang", "Yujun Cai", "Zhicheng Yang", "Jing Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3884b1807cbee971364947aa2f3756937c4c774d9942bb4da3d0fabb69bda07", "submitted_at": "2025-06-18T12:55:35Z", "updated_at": "2025-06-18T12:55:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.15425v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.15425", "identifiers": {"arxiv": "2506.15425"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.15425"}], "name": "Understanding GUI Agent Localization Biases through Logit Sharpness", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a finer-grained taxonomy of GUI localization errors and a logit-based Peak Sharpness Score for analyzing coordinate-prediction uncertainty. A context-aware cropping method refines the visual input without retraining the grounding model.", "topics": ["grounding", "evaluation", "robustness", "training-free"], "updated_at": "2025-06-18", "url": "https://arxiv.org/abs/2506.15425"}
{"aliases": [], "authors": ["Nikolay Blagoev", "Oğuzhan Ersoy", "Lydia Yiyu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a73ca5abcbaa032bdef033d9f1f98f37978a2bd53fd495b1ba079133909351bc", "submitted_at": "2025-06-18T13:48:33Z", "updated_at": "2026-04-04T20:02:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.15461v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.15461", "identifiers": {"arxiv": "2506.15461"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.15461"}], "name": "All is Not Lost: LLM Recovery without Checkpoints", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-04", "url": "https://arxiv.org/abs/2506.15461"}
{"aliases": [], "authors": ["Yining Hong", "Rui Sun", "Bingxuan Li", "Xingcheng Yao", "Maxine Wu", "Alexander Chien", "Da Yin", "Ying Nian Wu", "Zhecan James Wang", "Kai-Wei Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e8f714b234a1aad60d0306df23c90410aebfd257fb5906644d0382479e8f3d0", "submitted_at": "2025-06-18T17:58:17Z", "updated_at": "2025-07-29T22:40:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.15677v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.15677", "identifiers": {"arxiv": "2506.15677"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.15677"}], "name": "Embodied Web Agents: Bridging Physical-Digital Realms for Integrated Agent Intelligence", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The physical component is a simulation; the supported catalog connection is its functional web interaction.", "published_at": "2025-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces environments that couple 3D physical-world simulations with functional web interfaces. The accompanying benchmark evaluates tasks requiring agents to coordinate online information and physical actions across cooking, navigation and shopping scenarios.", "topics": ["evaluation", "web-navigation", "planning"], "updated_at": "2025-07-29", "url": "https://arxiv.org/abs/2506.15677"}
{"aliases": [], "authors": ["Reyna Abhyankar", "Qi Qi", "Yiying Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df3f654ef1364aeb6cdaeaea1c6334da949a798b2f8256c9ddc4e6a88a588e19", "submitted_at": "2025-06-19T05:26:40Z", "updated_at": "2026-05-18T17:55:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.16042v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.16042", "identifiers": {"arxiv": "2506.16042"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.16042"}], "name": "OSWorld-Human: Benchmarking the Efficiency of Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "OSWorld-Human adds human-determined trajectories to OSWorld to evaluate the efficiency of desktop computer-use agents. The study compares 16 agents and analyzes model-call latency, changes in step duration during task execution, and excess action steps.", "topics": ["evaluation", "efficiency", "observability"], "updated_at": "2026-05-18", "url": "https://arxiv.org/abs/2506.16042"}
{"aliases": [], "authors": ["Atharv Singh Patlan", "Ashwin Hebbar", "Pramod Viswanath", "Prateek Mittal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d346389013d465f578417a7fca80ee527f5e7b582e93eb73dae148f11479971", "submitted_at": "2025-06-18T14:29:02Z", "updated_at": "2025-06-18T14:29:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.17318v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.17318", "identifiers": {"arxiv": "2506.17318"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.17318"}], "name": "Context manipulation attacks : Web agents are susceptible to corrupted memory", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies context manipulation in web agents by altering stored task representations rather than webpage prompts alone. Evaluation on Browser-use and Agent-E examines memory-directed attacks and the implications for secure context handling.", "topics": ["safety-security", "memory", "evaluation"], "updated_at": "2025-06-18", "url": "https://arxiv.org/abs/2506.17318"}
{"aliases": [], "authors": ["Jinjie Wei", "Jiyao Liu", "Lihao Liu", "Ming Hu", "Junzhi Ning", "Mingcheng Li", "Weijie Yin", "Junjun He", "Xiao Liang", "Chao Feng", "Dingkang Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f12d2ae58243b5f89a6a510e8f3ef4073a605c4bafdf63c37fd38db57ea84602", "submitted_at": "2025-06-22T06:30:52Z", "updated_at": "2026-07-21T05:30:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.17913v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.17913", "identifiers": {"arxiv": "2506.17913"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.17913"}], "name": "Learning, Reasoning, Refinement: A Framework for Kahneman's Dual-System Intelligence in GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "CogniGUI combines hierarchical visual parsing of interface elements with a GRPO-based grounding agent that compares alternative interaction paths. Its ScreenSeek benchmark evaluates navigation across applications, changing interface states, and consistency between interfaces.", "topics": ["grounding", "planning", "reinforcement-learning", "evaluation"], "updated_at": "2026-07-21", "url": "https://arxiv.org/abs/2506.17913"}
{"aliases": [], "authors": ["Xinzge Gao", "Chuanrui Hu", "Bin Chen", "Teng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aaa62e24a7a269fea49471f17028a3d5d8068d934e450f53bc002ae80fb7cb03", "submitted_at": "2025-06-22T20:17:46Z", "updated_at": "2025-06-22T20:17:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.18158v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.18158", "identifiers": {"arxiv": "2506.18158"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.18158"}], "name": "Chain-of-Memory: Enhancing GUI Agents for Cross-Application Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Chain-of-Memory explicitly records action descriptions and task-relevant screen information in short- and long-term memory for GUI navigation. GUI Odyssey-CoM provides 111,000 screen-action pairs with memory annotations for learning and evaluating information retention across applications.", "topics": ["memory", "state-tracking", "cross-app", "data-generation"], "updated_at": "2025-06-22", "url": "https://arxiv.org/abs/2506.18158"}
{"aliases": [], "authors": ["Shuangbao Paul Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0b30a8448849358dee8a24a0a82f6dc98df8f3bab02209e3d6b4000cd1c0d1f3", "submitted_at": "2025-06-23T15:49:20Z", "updated_at": "2025-06-23T15:49:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.18780v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.18780", "identifiers": {"arxiv": "2506.18780"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.18780"}], "name": "Design high-confidence computers using trusted instructional set architecture and emulators", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-23", "url": "https://arxiv.org/abs/2506.18780"}
{"aliases": [], "authors": ["Weizhi Zhang", "Yangning Li", "Yuanchen Bei", "Junyu Luo", "Guancheng Wan", "Liangwei Yang", "Chenxuan Xie", "Yuyao Yang", "Wei-Chieh Huang", "Chunyu Miao", "Henry Peng Zou", "Xiao Luo", "Yusheng Zhao", "Yankai Chen", "Chunkit Chan", "Peilin Zhou", "Xinyang Zhang", "Chenwei Zhang", "Jingbo Shang", "Ming Zhang", "Yangqiu Song", "Irwin King", "Philip S. Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fda1759e7e120089a27129c6e76e0cf5702a9afcf4d254c835f27fce2e99fb17", "submitted_at": "2025-06-23T17:27:19Z", "updated_at": "2025-07-03T17:48:36Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.18959v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.18959", "identifiers": {"arxiv": "2506.18959"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.18959"}], "name": "From Web Search towards Agentic Deep Research: Incentivizing Search with Reasoning Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-03", "url": "https://arxiv.org/abs/2506.18959"}
{"aliases": [], "authors": ["Aleksandr Algazinov", "Matt Laing", "Paul Laban"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "615cc74d3dc42e042e07853f0528cadc9ab46b01dbcfaeadbfc659910415fa2f", "submitted_at": "2025-06-24T10:40:23Z", "updated_at": "2025-07-15T06:04:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.19502v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.19502", "identifiers": {"arxiv": "2506.19502"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.19502"}], "name": "MATE: LLM-Powered Multi-Agent Translation Environment for Accessibility Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-15", "url": "https://arxiv.org/abs/2506.19502"}
{"aliases": [], "authors": ["Jihao Gu", "Qihang Ai", "Yingyao Wang", "Pi Bu", "Jingxuan Xing", "Zekun Zhu", "Wei Jiang", "Ziming Wang", "Yingxiu Zhao", "Ming-Liang Zhang", "Jun Song", "Yuning Jiang", "Bo Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9620eb7d8fa5f5c2a318ff516d45b8a4cfd35cf0c235b96bab3bba4e5a6aab9b", "submitted_at": "2025-06-25T11:34:43Z", "updated_at": "2026-04-25T13:54:32Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2506.20332v4"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2506.20332", "identifiers": {"arxiv": "2506.20332"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.20332"}], "name": "Mobile-R1: Towards Interactive Capability for VLM-Based Mobile Agent via Systematic Training", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses a staged curriculum from response-format alignment through action feedback to interactive task-level training for mobile agents.", "topics": ["reinforcement-learning", "mobile-navigation"], "updated_at": "2026-04-25", "url": "https://arxiv.org/abs/2506.20332"}
{"aliases": [], "authors": ["Sungjae Park", "Homanga Bharadhwaj", "Shubham Tulsiani"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2919a5f08d83a5b74dfd7aeab4f800c6f962440abf445b185a9eb3ff90611ae9", "submitted_at": "2025-06-25T17:59:01Z", "updated_at": "2026-06-12T19:46:40Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.20668v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.20668", "identifiers": {"arxiv": "2506.20668"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.20668"}], "name": "DemoDiffusion: One-Shot Human Imitation using pre-trained Diffusion Policy", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-12", "url": "https://arxiv.org/abs/2506.20668"}
{"aliases": [], "authors": ["Tianyi Men", "Zhuoran Jin", "Pengfei Cao", "Yubo Chen", "Kang Liu", "Jun Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b251895e2b99f01d6a80038fa5b99c5bab17ee6c38fd515c5ef00244575b2525", "submitted_at": "2025-06-26T13:36:12Z", "updated_at": "2025-06-26T13:36:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.21252v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.21252", "identifiers": {"arxiv": "2506.21252"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.21252"}], "name": "Agent-RewardBench: Towards a Unified Benchmark for Reward Modeling across Perception, Planning, and Safety in Real-World Multimodal Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-26", "url": "https://arxiv.org/abs/2506.21252"}
{"aliases": [], "authors": ["Jack Lanchantin", "Angelica Chen", "Janice Lan", "Xian Li", "Swarnadeep Saha", "Tianlu Wang", "Jing Xu", "Ping Yu", "Weizhe Yuan", "Jason E Weston", "Sainbayar Sukhbaatar", "Ilia Kulikov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4a6a870941e8c334cd8f5b35cbed3536a07fcf928b0425172a10dccd4002e3a6", "submitted_at": "2025-06-26T17:25:49Z", "updated_at": "2025-06-26T17:25:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.21495v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.21495", "identifiers": {"arxiv": "2506.21495"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.21495"}], "name": "Bridging Offline and Online Reinforcement Learning for LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-06-26", "url": "https://arxiv.org/abs/2506.21495"}
{"aliases": [], "authors": ["Boyu Gou", "Zanming Huang", "Yuting Ning", "Yu Gu", "Michael Lin", "Weijian Qi", "Andrei Kopanev", "Botao Yu", "Bernal Jiménez Gutiérrez", "Yiheng Shu", "Chan Hee Song", "Jiaman Wu", "Shijie Chen", "Hanane Nour Moussa", "Tianshu Zhang", "Jian Xie", "Yifei Li", "Tianci Xue", "Zeyi Liao", "Kai Zhang", "Boyuan Zheng", "Zhaowei Cai", "Viktor Rozgic", "Morteza Ziyadi", "Huan Sun", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0b614feb7733b9d8b287e370d7b61c491b51ef935e46451719002c85dc9ed6f", "submitted_at": "2025-06-26T17:32:50Z", "updated_at": "2025-07-03T15:47:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.21506v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T08:01:28Z", "role": "primary", "sha256": "822cec31fa0db7ffa390fda806d88c2f9c0fe4d328742e7008b662916e92373a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2005.07648%2C2206.04615%2C2311.12022%2C2311.16922%2C2502.13130%2C2506.21506%2C2511.11332&max_results=7"}], "id": "arxiv:2506.21506", "identifiers": {"arxiv": "2506.21506"}, "kind": "paper", "last_verified_at": "2026-10-18T08:01:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.21506"}], "name": "Mind2Web 2: Evaluating Agentic Search with Agent-as-a-Judge", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-03", "url": "https://arxiv.org/abs/2506.21506"}
{"aliases": [], "authors": ["Yixiong Fang", "Tianran Sun", "Yuling Shi", "Min Wang", "Xiaodong Gu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd75d10b8e20d8b2961d965ab05fd6f4dd52994ea9c5079c40081e6024df0196", "submitted_at": "2025-06-21T13:01:04Z", "updated_at": "2025-09-14T14:29:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.21614v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2506.21614", "identifiers": {"arxiv": "2506.21614"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2506.21614"}], "name": "LastingBench: Defend Benchmarks Against Knowledge Leakage", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-06-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-14", "url": "https://arxiv.org/abs/2506.21614"}
{"aliases": [], "authors": ["ZongHan Hsieh", "Tzer-Jen Wei", "ShengJing Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb3678b0d0e645a87ce9392b606c1d67dcbc40656880c57d1cbb8ddd8ede7ea1", "submitted_at": "2025-06-30T03:33:02Z", "updated_at": "2025-07-18T06:03:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.23491v3"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2506.23491", "identifiers": {"arxiv": "2506.23491"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.23491"}], "name": "ZonUI-3B: A Lightweight Vision-Language Model for Cross-Resolution GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "ZonUI-3B trains GUI grounding with a curated 24,000-example collection of mobile, desktop, and web screenshots. Its two-stage procedure first learns across platforms and then specializes on high-resolution images, with ablations on data diversity and redundant examples.", "topics": ["grounding", "data-generation", "efficiency"], "updated_at": "2025-07-18", "url": "https://arxiv.org/abs/2506.23491"}
{"aliases": [], "authors": ["Hang Wu", "Hongkai Chen", "Yujun Cai", "Chang Liu", "Qingwen Ye", "Ming-Hsuan Yang", "Yiwei Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e423834b07e74155214597c0fbb66c26b949e2b726568a65617a670a9ab157fa", "submitted_at": "2025-06-12T03:13:21Z", "updated_at": "2025-09-05T17:21:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.00008v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.00008", "identifiers": {"arxiv": "2507.00008"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.00008"}], "name": "DiMo-GUI: Advancing Test-time Scaling in GUI Grounding via Modality-Aware Visual Reasoning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates textual and iconic UI elements and iteratively focuses visual regions to refine grounding without additional model training.", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-09-05", "url": "https://arxiv.org/abs/2507.00008"}
{"aliases": [], "authors": ["Imene Kerboua", "Sahar Omidi Shayegan", "Megh Thakkar", "Xing Han Lù", "Massimo Caccia", "Véronique Eglin", "Alexandre Aussem", "Jérémy Espinas", "Alexandre Lacoste"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b8972688fb95908c024a85e3ffc258be61280ee8d22e6ab97dfb0c4997b83445", "submitted_at": "2025-06-30T19:24:45Z", "updated_at": "2025-06-30T19:24:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.00210v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.00210", "identifiers": {"arxiv": "2507.00210"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.00210"}], "name": "LineRetriever: Planning-Aware Observation Reduction for Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-06-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "LineRetriever uses a language model to select observation lines that matter for upcoming web-navigation actions. It reduces DOM or accessibility-tree context by considering the planning horizon instead of relying solely on semantic similarity.", "topics": ["web-navigation", "efficiency", "planning"], "updated_at": "2025-06-30", "url": "https://arxiv.org/abs/2507.00210"}
{"aliases": [], "authors": ["Siyuan Liang", "Tianmeng Fang", "Zhe Liu", "Aishan Liu", "Yan Xiao", "Jinyuan He", "Ee-Chien Chang", "Xiaochun Cao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "78dd523ad6123fe1ee5164685a3853719c9a4442adac8661050f147ff6a9f619", "submitted_at": "2025-07-01T15:10:00Z", "updated_at": "2025-07-01T15:10:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.00841v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:16.030926+00:00", "role": "primary", "sha256": "4dcd1ba66ba21c0f2cb2aeb00c8bcab7c0cee4da37ff2fa6c16b45066aaa9db0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:16.030926+00:00", "role": "primary", "sha256": "4dcd1ba66ba21c0f2cb2aeb00c8bcab7c0cee4da37ff2fa6c16b45066aaa9db0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:10:56.912207+00:00", "role": "primary", "sha256": "a25da6700a238167d5e9915cf6fb3a4bcd910cefc300f4be6eed56271be26f00", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2507.00841v1"}], "id": "arxiv:2507.00841", "identifiers": {"arxiv": "2507.00841"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:16.030926+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.00841"}], "name": "SafeMobile: Chain-level Jailbreak Detection and Automated Evaluation for Multimodal Mobile Agents", "notes": "The reviewed evaluation uses Android emulation and a specific MobileSafetyBench-derived setup. Its judge and score threshold are evaluator choices, not universal safety measurements, and its task population is not the whole later MobileSafetyBench corpus. No results were reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (II.B Defense objective; III Chain-level guard and GPTJudge; IV.B Evaluation setup; V.D Judge validity); not an entire-paper reading or reproduction claim.", "published_at": "2025-07-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies chain-level risk detection for mobile multimodal agents and a language-model-assisted evaluator, assessing safety interventions together with functional preservation.", "topics": ["safety-security", "evaluation", "mobile-navigation"], "updated_at": "2025-07-01", "url": "https://arxiv.org/abs/2507.00841"}
{"aliases": [], "authors": ["Zihao Sun", "Zijing Shi", "Ling Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1745ec9cf10c465c0f389f885df6db502f1884cca271bccea1b45acd2c787af9", "submitted_at": "2025-07-01T16:43:57Z", "updated_at": "2026-09-24T12:58:22Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.00938v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.00938", "identifiers": {"arxiv": "2507.00938"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.00938"}], "name": "WebArxiv: A Reproducible Benchmark for Evaluating Multimodal Web Agents on arXiv Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebArxiv provides 510 static-snapshot tasks with deterministic answers for scholarly browsing, including constrained paper retrieval, content extraction, and cross-paper comparison. The study also tests dynamic retrieval of interaction history to address repetitive or incomplete agent reasoning.", "topics": ["evaluation", "web-navigation", "memory"], "updated_at": "2026-09-24", "url": "https://arxiv.org/abs/2507.00938"}
{"aliases": [], "authors": ["Zhaoyan Sun", "Jiayi Wang", "Xinyang Zhao", "Jiachi Wang", "Guoliang Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ec8fcae2fa421b0a81900313c84109da1f82465c70afb59e17f5fb3107c21565", "submitted_at": "2025-07-02T11:04:49Z", "updated_at": "2025-07-02T11:04:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.01599v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.01599", "identifiers": {"arxiv": "2507.01599"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.01599"}], "name": "Data Agent: A Holistic Architecture for Orchestrating Data+AI Ecosystems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-02", "url": "https://arxiv.org/abs/2507.01599"}
{"aliases": [], "authors": ["Juntao Liu", "Liqiang Niu", "Wenchao Chen", "Jie Zhou", "Fandong Meng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "901ae56085a1370c0cca10896bceb76766a44c1b97a5fa759b72bd0e612e2ea1", "submitted_at": "2025-07-03T03:42:54Z", "updated_at": "2025-07-03T03:42:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.02279v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.02279", "identifiers": {"arxiv": "2507.02279"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.02279"}], "name": "LaCo: Efficient Layer-wise Compression of Visual Tokens for Multimodal Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-03", "url": "https://arxiv.org/abs/2507.02279"}
{"aliases": [], "authors": ["Kuan Li", "Zhongwang Zhang", "Huifeng Yin", "Liwen Zhang", "Litu Ou", "Jialong Wu", "Wenbiao Yin", "Baixuan Li", "Zhengwei Tao", "Xinyu Wang", "Weizhou Shen", "Junkai Zhang", "Dingchu Zhang", "Xixi Wu", "Yong Jiang", "Ming Yan", "Pengjun Xie", "Fei Huang", "Jingren Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "101b0296b51a9c790e473f334145bf1f0914cd6ee7d3616cba9fd8ad74c8a7d1", "submitted_at": "2025-07-03T12:59:07Z", "updated_at": "2025-07-03T12:59:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.02592v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.02592", "identifiers": {"arxiv": "2507.02592"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.02592"}], "name": "WebSailor: Navigating Super-human Reasoning for Web Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-03", "url": "https://arxiv.org/abs/2507.02592"}
{"aliases": [], "authors": ["Yuxuan Zhu", "Tengjun Jin", "Yada Pruksachatkun", "Andy Zhang", "Shu Liu", "Sasha Cui", "Sayash Kapoor", "Shayne Longpre", "Kevin Meng", "Rebecca Weiss", "Fazl Barez", "Rahul Gupta", "Jwala Dhamala", "Jacob Merizian", "Mario Giulianelli", "Harry Coppock", "Cozmin Ududec", "Jasjeet Sekhon", "Jacob Steinhardt", "Antony Kellermann", "Sarah Schwettmann", "Matei Zaharia", "Ion Stoica", "Percy Liang", "Daniel Kang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a55c6ccded8854701c0f07a235cd6b5021d99f33df64faf457329f3f317f8ea", "submitted_at": "2025-07-03T17:35:31Z", "updated_at": "2025-08-07T06:58:08Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2507.02825v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.02825", "identifiers": {"arxiv": "2507.02825"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.02825"}], "name": "Establishing Best Practices for Building Rigorous Agentic Benchmarks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-07", "url": "https://arxiv.org/abs/2507.02825"}
{"aliases": [], "authors": ["Gongwei Chen", "Xurui Zhou", "Rui Shao", "Yibo Lyu", "Kaiwen Zhou", "Shuai Wang", "Wentao Li", "Yinchuan Li", "Zhongang Qi", "Liqiang Nie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "774e2515b563f75d4561daaa6a4b672423d917a3f50e98308ab448e2ae0731e5", "submitted_at": "2025-07-04T17:37:15Z", "updated_at": "2025-07-04T17:37:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.03730v1"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.03730", "identifiers": {"arxiv": "2507.03730"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.03730"}], "name": "Less is More: Empowering GUI Agent with Context-Aware Simplification", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "SimpAgent prunes distracting interface elements with visual masks and compresses interaction history using consistency-based guidance. The study evaluates these two forms of context reduction during GUI navigation in web and mobile environments.", "topics": ["efficiency", "interaction-history", "web-navigation", "mobile-navigation"], "updated_at": "2025-07-04", "url": "https://arxiv.org/abs/2507.03730"}
{"aliases": [], "authors": ["Jingwei Shi", "Zeyu Zhang", "Biao Wu", "Yanjie Liang", "Meng Fang", "Ling Chen", "Yang Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5eb51028622be44857197c0a83de60d6adf5daba46d15884b2bc47509a3bddcc", "submitted_at": "2025-07-05T13:24:15Z", "updated_at": "2025-07-05T13:24:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.04036v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.04036", "identifiers": {"arxiv": "2507.04036"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.04036"}], "name": "PresentAgent: Multimodal Agent for Presentation Video Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-05", "url": "https://arxiv.org/abs/2507.04036"}
{"aliases": [], "authors": ["Dheeraj Vattikonda", "Santhoshi Ravichandran", "Emiliano Penaloza", "Hadi Nekoei", "Megh Thakkar", "Thibault Le Sellier de Chezelles", "Nicolas Gontier", "Miguel Muñoz-Mármol", "Sahar Omidi Shayegan", "Stefania Raimondo", "Xue Liu", "Alexandre Drouin", "Laurent Charlin", "Alexandre Piché", "Alexandre Lacoste", "Massimo Caccia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15fa527a869ab66f4f39900ab07ea8b8db930703a8e950ed1e195e1e66164099", "submitted_at": "2025-07-05T17:12:33Z", "updated_at": "2026-02-13T17:24:17Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.04103v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.04103", "identifiers": {"arxiv": "2507.04103"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.04103"}], "name": "How to Train Your LLM Web Agent: A Statistical Diagnosis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies compute allocation for web-agent post-training with teacher-generated supervised fine-tuning followed by on-policy reinforcement learning. It samples 1,370 training configurations and uses bootstrap analysis to examine hyperparameter sensitivity on WorkArena and MiniWoB++.", "topics": ["imitation-learning", "reinforcement-learning", "efficiency", "evaluation"], "updated_at": "2026-02-13", "url": "https://arxiv.org/abs/2507.04103"}
{"aliases": [], "authors": ["Guohong Liu", "Jialei Ye", "Jiacheng Liu", "Yuanchun Li", "Wei Liu", "Pengzhi Gao", "Jian Luan", "Yunxin Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "85ebf8b3b5916e66e4b4221c73bb29519156c41e8d1695bcc46d275dd50cf58d", "submitted_at": "2025-07-06T03:31:36Z", "updated_at": "2026-04-14T14:47:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.04227v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.04227", "identifiers": {"arxiv": "2507.04227"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.04227"}], "name": "Mobile GUI Agents under Real-world Threats: Are We There Yet?", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-07-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces an app-content instrumentation framework and complementary interactive and static test suites for mobile GUI agents. It studies how untrusted third-party content changes task execution and state understanding.", "topics": ["mobile-navigation", "safety-security", "evaluation"], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2507.04227"}
{"aliases": [], "authors": ["Yuanzhe Hu", "Yu Wang", "Julian McAuley"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0819dc2369db97e944f81981594f662b6b556bfaa79cd6b98be03a6807a4e90c", "submitted_at": "2025-07-07T17:59:54Z", "updated_at": "2026-06-28T17:25:01Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.05257v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2507.05257", "identifiers": {"arxiv": "2507.05257"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.05257"}], "name": "Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-28", "url": "https://arxiv.org/abs/2507.05257"}
{"aliases": [], "authors": ["Daniel Jones", "Giorgio Severi", "Martin Pouliot", "Gary Lopez", "Joris de Gruyter", "Santiago Zanella-Beguelin", "Justin Song", "Blake Bullwinkel", "Pamela Cortez", "Amanda Minnich"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3ad2f299303076e3474cf247cf25d20ba63b1f1aa8601ab74243c00896bb56e4", "submitted_at": "2025-07-07T19:50:21Z", "updated_at": "2025-07-07T19:50:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.05445v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.05445", "identifiers": {"arxiv": "2507.05445"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.05445"}], "name": "A Systematization of Security Vulnerabilities in Computer Use Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. This is a threat analysis with case studies; the annotation does not imply a standardized public benchmark or independently validated defenses.", "published_at": "2025-07-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes security boundaries in computer-use agents through a risk taxonomy and concrete adversarial case studies. It links observed failures to input provenance, interface-action binding, memory, and delegation, and proposes CUA-specific evaluation principles.", "topics": ["safety-security", "evaluation"], "updated_at": "2025-07-07", "url": "https://arxiv.org/abs/2507.05445"}
{"aliases": [], "authors": ["Joonhyung Park", "Peng Tang", "Sagnik Das", "Srikar Appalaraju", "Kunwar Yashraj Singh", "R. Manmatha", "Shabnam Ghadar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aff9b208c446bd57f24258734557e0cd5caaf9146af68e594ec6a7e575919a6f", "submitted_at": "2025-07-08T04:56:57Z", "updated_at": "2025-07-08T04:56:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.05673v1"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.05673", "identifiers": {"arxiv": "2507.05673"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.05673"}], "name": "R-VLM: Region-Aware Vision Language Model for Precise GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "R-VLM localizes GUI elements through zoomed-in region proposals and an intersection-over-union-aware training objective. It evaluates the grounding model on ScreenSpot and AgentStudio and studies its effect on navigation in AITW and Mind2Web.", "topics": ["grounding", "screen-understanding", "web-navigation"], "updated_at": "2025-07-08", "url": "https://arxiv.org/abs/2507.05673"}
{"aliases": [], "authors": ["Yucheng Shi", "Wenhao Yu", "Zaitang Li", "Yonglin Wang", "Hongming Zhang", "Ninghao Liu", "Haitao Mi", "Dong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45d36dd32b30bf4faee6d9a72e3cdfe91edff465bbd886bdfc3c27c3da475b82", "submitted_at": "2025-07-08T07:07:53Z", "updated_at": "2025-07-08T07:07:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.05720v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.05720", "identifiers": {"arxiv": "2507.05720"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.05720"}], "name": "MobileGUI-RL: Advancing Mobile GUI Agent through Reinforcement Learning in Online Environment", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileGUI-RL constructs an online training curriculum through agent exploration and task filtering. It adapts GRPO with trajectory-aware advantages and rewards for task completion and execution efficiency, evaluated on three online mobile-agent benchmarks.", "topics": ["mobile-navigation", "reinforcement-learning", "data-generation", "efficiency"], "updated_at": "2025-07-08", "url": "https://arxiv.org/abs/2507.05720"}
{"aliases": [], "authors": ["Yan Yang", "Dongxu Li", "Yutong Dai", "Yuhao Yang", "Ziyang Luo", "Zirui Zhao", "Zhiyuan Hu", "Junzhe Huang", "Amrita Saha", "Zeyuan Chen", "Ran Xu", "Liyuan Pan", "Silvio Savarese", "Caiming Xiong", "Junnan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c1f659a30e1a0bd04219b9383927ce72e5f5b0d20783a89b850036d087155a4", "submitted_at": "2025-07-08T08:52:18Z", "updated_at": "2025-10-03T23:50:19Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2507.05791v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2507.05791", "identifiers": {"arxiv": "2507.05791"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.05791"}], "name": "GTA1: GUI Test-time Scaling Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "GTA1 samples multiple action proposals at each GUI step and uses a judge to select among them. A separate grounding model is trained with reinforcement learning to map selected proposals to visual targets, linking planning-time search with click localization.", "topics": ["planning", "grounding", "reinforcement-learning"], "updated_at": "2025-10-03", "url": "https://arxiv.org/abs/2507.05791"}
{"aliases": [], "authors": ["Sanidhya Vijayvargiya", "Aditya Bharat Soni", "Xuhui Zhou", "Zora Zhiruo Wang", "Nouha Dziri", "Graham Neubig", "Maarten Sap"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db56d120a7f04f9181a107bcd75c61a3eb3ac36939e2a0517d622332c70f3eac", "submitted_at": "2025-07-08T16:18:54Z", "updated_at": "2026-02-16T22:34:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.06134v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2507.06134", "identifiers": {"arxiv": "2507.06134"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.06134"}], "name": "OpenAgentSafety: A Comprehensive Framework for Evaluating Real-World AI Agent Safety", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-07-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates agent safety across real browser, code, filesystem, shell, and messaging tools using rule-based and model-based assessments.", "topics": ["safety-security", "evaluation"], "updated_at": "2026-02-16", "url": "https://arxiv.org/abs/2507.06134"}
{"aliases": [], "authors": ["Ziang Ye", "Yang Zhang", "Wentao Shi", "Xiaoyu You", "Fuli Feng", "Tat-Seng Chua"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "010a3a54ba37417c97e52434211375447418dd2e7ce1da6c339532534f0a920e", "submitted_at": "2025-07-09T14:36:00Z", "updated_at": "2025-09-24T14:33:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.06899v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.06899", "identifiers": {"arxiv": "2507.06899"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.06899"}], "name": "VisualTrap: A Stealthy Backdoor Attack on GUI Agents via Visual Grounding Manipulation", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-07-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies backdoors that redirect GUI visual grounding even when the task plan is correct. VisualTrap evaluates whether poisoned grounding training transfers to downstream tasks and across web, mobile, and desktop interface settings.", "topics": ["safety-security", "grounding", "evaluation"], "updated_at": "2025-09-24", "url": "https://arxiv.org/abs/2507.06899"}
{"aliases": [], "authors": ["Zaifeng Pan", "Ajjkumar Patel", "Zhengding Hu", "Yipeng Shen", "Yue Guan", "Wan-Lu Li", "Lianhui Qin", "Yida Wang", "Yufei Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb81df4795bb1ec842897d65866ad93be0274ddaf04e8769adcf7c1dd7f8e667", "submitted_at": "2025-07-10T03:39:23Z", "updated_at": "2025-07-10T03:39:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.07400v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.07400", "identifiers": {"arxiv": "2507.07400"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.07400"}], "name": "KVFlow: Efficient Prefix Caching for Accelerating LLM-Based Multi-Agent Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-10", "url": "https://arxiv.org/abs/2507.07400"}
{"aliases": [], "authors": ["Yuheng Huang", "Jiayang Song", "Da Song", "Zhenlan Ji", "Wenhan Wang", "Shuai Wang", "Lei Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b83eb40eb9fa4dcaeba000111982e31fa6fc1eeb517cf2f7ee61d7d70053faf1", "submitted_at": "2025-07-13T03:52:51Z", "updated_at": "2025-12-02T16:25:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.09481v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.09481", "identifiers": {"arxiv": "2507.09481"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.09481"}], "name": "Evaluating LLMs on Sequential API Call Through Automated Test Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-02", "url": "https://arxiv.org/abs/2507.09481"}
{"aliases": [], "authors": ["Xuzhao Li", "Xuchen Li", "Shiyu Hu", "Yongzhen Guo", "Wentao Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8dc1a5e14b7d7868acfc7796fe5a07980e0b5e092967407a3b155f7f8fd7607d", "submitted_at": "2025-07-14T03:45:24Z", "updated_at": "2025-07-26T11:17:08Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2507.09884v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.09884", "identifiers": {"arxiv": "2507.09884"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.09884"}], "name": "VerifyBench: A Systematic Benchmark for Evaluating Reasoning Verifiers Across Domains", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-26", "url": "https://arxiv.org/abs/2507.09884"}
{"aliases": [], "authors": ["Zihe Yan", "Jiaping Gui", "Zhuosheng Zhang", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5db0008a053357b6dba38b55f29642ddbcbf38761c64684fd1725183ff8e56d3", "submitted_at": "2025-07-13T08:36:09Z", "updated_at": "2026-04-07T11:46:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2507.10610v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.10610", "identifiers": {"arxiv": "2507.10610"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.10610"}], "name": "LaSM: Layer-wise Scaling Mechanism for Defending Pop-up Attack on GUI Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-07-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces LaSM, a training-free defense that selectively scales attention and MLP components to improve alignment with task-relevant screen regions. It evaluates the approach against pop-up-based GUI injection attacks.", "topics": ["safety-security", "grounding", "training-free"], "updated_at": "2026-04-07", "url": "https://arxiv.org/abs/2507.10610"}
{"aliases": [], "authors": ["Joel Nitu", "Heidrun Mühle", "Andreas Stöckl"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9fe2d9d877a0666d46a065bd8b673c282c879ef0a2620e3b267793a91858b519", "submitted_at": "2025-07-17T07:10:27Z", "updated_at": "2025-11-12T13:34:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.12844v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.12844", "identifiers": {"arxiv": "2507.12844"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.12844"}], "name": "Machine-Readable Ads: Accessibility and Trust Patterns for AI Web Agents interacting with Online Advertisements", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-07-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares DOM-based and pixel-based web agents on a cloned news site containing advertisements, consent dialogs and paywalls. The study examines how interface presentation affects ad interaction and consequential subscription choices.", "topics": ["evaluation", "ui-design", "safety-security"], "updated_at": "2025-11-12", "url": "https://arxiv.org/abs/2507.12844"}
{"aliases": [], "authors": ["Senqiao Yang", "Junyi Li", "Xin Lai", "Bei Yu", "Hengshuang Zhao", "Jiaya Jia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "807a70c7f87af366ecd7aa53c058916d0f735d652761f6530cdd5835032ca995", "submitted_at": "2025-07-17T17:59:55Z", "updated_at": "2025-07-17T17:59:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.13348v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.13348", "identifiers": {"arxiv": "2507.13348"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.13348"}], "name": "VisionThink: Smart and Efficient Vision Language Model via Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-17", "url": "https://arxiv.org/abs/2507.13348"}
{"aliases": [], "authors": ["Boyuan Zheng", "Zeyi Liao", "Scott Salisbury", "Zeyuan Liu", "Michael Lin", "Qinyuan Zheng", "Zifan Wang", "Xiang Deng", "Dawn Song", "Huan Sun", "Yu Su"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8f082fd40174e01f40e97d64f7aa3d90e5f76e831df8f93afdc21aa2e35bfd5", "submitted_at": "2025-07-18T18:06:27Z", "updated_at": "2025-07-18T18:06:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.14293v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.14293", "identifiers": {"arxiv": "2507.14293"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.14293"}], "name": "WebGuard: Building a Generalizable Guardrail for Web Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-07-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebGuard, a human-annotated dataset for predicting outcomes and risk levels of state-changing website actions. It provides generalization splits and evaluates specialized guardrail models.", "topics": ["safety-security", "reward-models", "evaluation"], "updated_at": "2025-07-18", "url": "https://arxiv.org/abs/2507.14293"}
{"aliases": [], "authors": ["Guancheng Zeng", "Xueyi Chen", "Jiawang Hu", "Shaohua Qi", "Yaxuan Mao", "Zhantao Wang", "Yifan Nie", "Shuang Li", "Qiuyang Feng", "Pengxu Qiu", "Yujia Wang", "Wenqiang Han", "Linyan Huang", "Gang Li", "Jingjing Mo", "Haowen Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d63dfeb18def57256964d19e9e1c413ff9c71ef1c2fd6c368884e20fd489a9a", "submitted_at": "2025-07-19T02:46:19Z", "updated_at": "2025-07-22T10:01:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.14447v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.14447", "identifiers": {"arxiv": "2507.14447"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.14447"}], "name": "Routine: A Structural Planning Framework for LLM Agent System in Enterprise", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-22", "url": "https://arxiv.org/abs/2507.14447"}
{"aliases": [], "authors": ["Sam Johnson", "Viet Pham", "Thai Le"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "79963ac401384f0c23ecc65d58e616daacc93103456ceb39241dcd6d71c6591c", "submitted_at": "2025-07-20T03:10:13Z", "updated_at": "2025-07-20T03:10:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.14799v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.14799", "identifiers": {"arxiv": "2507.14799"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.14799"}], "name": "Manipulating LLM Web Agents with Indirect Prompt Injection Attack via HTML Accessibility Tree", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-07-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies indirect prompt injection through HTML content consumed as accessibility trees. The evaluation tests how manipulated page observations redirect browser-agent actions on real websites.", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2025-07-20", "url": "https://arxiv.org/abs/2507.14799"}
{"aliases": [], "authors": ["Fei Tang", "Zhangxuan Gu", "Zhengxi Lu", "Xuyang Liu", "Shuheng Shen", "Changhua Meng", "Wen Wang", "Wenqi Zhang", "Yongliang Shen", "Weiming Lu", "Jun Xiao", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2a2a49373259cb5903fad8c2fb48a7380732642bedd81a644e735af4f028c63", "submitted_at": "2025-07-21T17:53:42Z", "updated_at": "2025-07-28T16:54:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2507.15846v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.15846", "identifiers": {"arxiv": "2507.15846"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.15846"}], "name": "GUI-G$^2$: Gaussian Reward Modeling for GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-G2 replaces binary grounding rewards with Gaussian point rewards and spatial-coverage rewards. An adaptive variance tied to element dimensions supplies continuous training feedback for targets of different sizes, evaluated on the ScreenSpot benchmark family.", "topics": ["grounding", "reinforcement-learning", "reward-models"], "updated_at": "2025-07-28", "url": "https://arxiv.org/abs/2507.15846"}
{"aliases": [], "authors": ["Viktor Muryn", "Marta Sumyk", "Mariya Hirna", "Sofiya Garkot", "Maksym Shamrai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9866105ab9e7ecb94d0f1388926f1faa8a6d959bc01128af68a84774f832e2e5", "submitted_at": "2025-07-22T15:38:12Z", "updated_at": "2025-07-22T15:38:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.16704v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.16704", "identifiers": {"arxiv": "2507.16704"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.16704"}], "name": "Screen2AX: Vision-Based Approach for Automatic macOS Accessibility Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-22", "url": "https://arxiv.org/abs/2507.16704"}
{"aliases": [], "authors": ["Ning Li", "Xiangmou Qu", "Jiamu Zhou", "Jun Wang", "Muning Wen", "Kounianhua Du", "Xingyu Lou", "Qiuying Peng", "Jun Wang", "Weinan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d58a188742fef5862b299c1dcf17bb6be139e81aafd3d2e1bdbe26f88001384", "submitted_at": "2025-07-21T09:37:05Z", "updated_at": "2025-07-21T09:37:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.16853v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.16853", "identifiers": {"arxiv": "2507.16853"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.16853"}], "name": "MobileUse: A GUI Agent with Hierarchical Reflection for Autonomous Mobile Operation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileUse reflects on errors at action and task timescales, invoking reflection when needed and using proactive exploration to learn unfamiliar apps. It evaluates this mobile execution framework on AndroidWorld and AndroidLab and describes a toolkit for physical-device operation.", "topics": ["mobile-navigation", "self-reflection", "robustness", "long-horizon"], "updated_at": "2025-07-21", "url": "https://arxiv.org/abs/2507.16853"}
{"aliases": [], "authors": ["Yuheng Zhao", "Xueli Shu", "Liwen Fan", "Lin Gao", "Yu Zhang", "Siming Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3c271f330c5bf27e2831efd96fb1d6f843bd79942791686a866df1b4812de2f", "submitted_at": "2025-07-24T08:02:35Z", "updated_at": "2025-07-24T08:02:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.18165v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:16.030926+00:00", "role": "primary", "sha256": "4dcd1ba66ba21c0f2cb2aeb00c8bcab7c0cee4da37ff2fa6c16b45066aaa9db0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:16.030926+00:00", "role": "primary", "sha256": "4dcd1ba66ba21c0f2cb2aeb00c8bcab7c0cee4da37ff2fa6c16b45066aaa9db0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:05:30.529800+00:00", "role": "primary", "sha256": "7ca9569113479791a0482f4c196002a8459f1826eb6a3c9fab8a701b33360593", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/html/2507.18165v1"}], "id": "arxiv:2507.18165", "identifiers": {"arxiv": "2507.18165"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:16.030926+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.18165"}], "name": "ProactiveVA: Proactive Visual Analytics with LLM-Based UI Agent", "notes": "Adjacent work on embedded assistance in two instrumented visual-analytics systems. Its in-lab user-study evidence is context-specific, and the paper discusses clarification and intervention-control trade-offs. General GUI-agent performance is not inferred. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (5.3 User study; 5.3.2 Observed issues; 6.1 Design trade-off; 6.2 Limitations); not an entire-paper reading or reproduction claim.", "published_at": "2025-07-24", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Studies proactive assistance in visual-analytics interfaces, using interaction logs to detect support needs and provide context-aware suggestions and guidance.", "topics": ["human-in-the-loop", "planning"], "updated_at": "2025-07-24", "url": "https://arxiv.org/abs/2507.18165"}
{"aliases": [], "authors": ["Xuetian Chen", "Yinghao Chen", "Xinfeng Yuan", "Zhuo Peng", "Lu Chen", "Yuekeng Li", "Zhoujia Zhang", "Yingqian Huang", "Leyan Huang", "Jiaqing Liang", "Tianbao Xie", "Zhiyong Wu", "Qiushi Sun", "Biqing Qi", "Bowen Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fcd5a34bf6187f48da8e08842678e1e21cd558b2bcc7aab2611ceee7bc0d0179", "submitted_at": "2025-07-25T10:14:53Z", "updated_at": "2025-07-25T10:14:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.19132v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.19132", "identifiers": {"arxiv": "2507.19132"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.19132"}], "name": "OS-MAP: How Far Can Computer-Using Agents Go in Breadth and Depth?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "OS-MAP organizes 416 computer-use tasks across 15 applications by a five-level automation taxonomy and a hierarchy of user demands. The benchmark separates required autonomy from generalization scope to diagnose where agents struggle as task demands increase.", "topics": ["evaluation", "automation", "planning"], "updated_at": "2025-07-25", "url": "https://arxiv.org/abs/2507.19132"}
{"aliases": [], "authors": ["Xuehui Wang", "Zhenyu Wu", "JingJing Xie", "Zichen Ding", "Bowen Yang", "Zehao Li", "Zhaoyang Liu", "Qingyun Li", "Xuan Dong", "Zhe Chen", "Weiyun Wang", "Xiangyu Zhao", "Jixuan Chen", "Haodong Duan", "Tianbao Xie", "Chenyu Yang", "Shiqian Su", "Yue Yu", "Yuan Huang", "Yiqian Liu", "Xiao Zhang", "Yanting Zhang", "Xiangyu Yue", "Weijie Su", "Xizhou Zhu", "Wei Shen", "Jifeng Dai", "Wenhai Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b91f7bacbac917e374aa6f656dcfbcb2e220ab3227fa2cff295c63a1c005f9f", "submitted_at": "2025-07-25T17:59:26Z", "updated_at": "2025-07-25T17:59:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.19478v1"}}, "environments": ["web", "desktop", "windows", "macos", "linux", "ios", "android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.19478", "identifiers": {"arxiv": "2507.19478"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.19478"}], "name": "MMBench-GUI: Hierarchical Multi-Platform Evaluation Framework for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "MMBench-GUI evaluates interface understanding, element grounding, task automation, and collaboration across six named desktop, mobile, and web platforms. It introduces the Efficiency-Quality Area metric for examining execution efficiency alongside task quality in online GUI automation.", "topics": ["evaluation", "grounding", "efficiency", "planning"], "updated_at": "2025-07-25", "url": "https://arxiv.org/abs/2507.19478"}
{"aliases": [], "authors": ["Sourena Khanzadeh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "269c2562915b462412e9f4d34f96932514260a9f3641b402cac81afc469a93fa", "submitted_at": "2025-07-26T10:10:02Z", "updated_at": "2025-07-26T10:10:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.19902v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.19902", "identifiers": {"arxiv": "2507.19902"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.19902"}], "name": "AgentMesh: A Cooperative Multi-Agent Generative AI Framework for Software Development Automation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-26", "url": "https://arxiv.org/abs/2507.19902"}
{"aliases": [], "authors": ["Qinglong Yang", "Haoming Li", "Haotian Zhao", "Xiaokai Yan", "Jingtao Ding", "Fengli Xu", "Yong Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "947c85efc189635a118b823b8ca0787ff6b1c922d1af8b21a8686b7df9a27cf8", "submitted_at": "2025-06-09T06:38:41Z", "updated_at": "2026-03-16T09:14:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.21071v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:11:36Z", "role": "primary", "sha256": "90a03448678a584f29ef11ae2afe1ef7040b4f3658c492f6e48eafe8fb6d2fa2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2507.21071", "identifiers": {"arxiv": "2507.21071"}, "kind": "paper", "last_verified_at": "2026-10-17T17:11:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.21071"}], "name": "FingerTip 20K: A Benchmark for Proactive and Personalized Mobile LLM Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "FingerTip 20K gathers 20K human Android demonstrations from users' long-term everyday app use, with user context, to evaluate proactive task suggestion and preference-aware task execution.", "topics": ["evaluation", "data-generation", "mobile-navigation", "interaction-history"], "updated_at": "2026-03-16", "url": "https://arxiv.org/abs/2507.21071"}
{"aliases": [], "authors": ["Yiming Xu", "Runzhi He", "Hengzhi Ye", "Minghui Zhou", "Huaimin Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "85634bb321e393aa7e742f2fe54257826cb71e8cca1909cbfb2a0efdd3d21322", "submitted_at": "2025-07-29T10:45:24Z", "updated_at": "2025-10-29T15:15:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.21678v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.21678", "identifiers": {"arxiv": "2507.21678"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.21678"}], "name": "Predicting Abandonment of Open Source Software Projects with An Integrated Feature Framework", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-29", "url": "https://arxiv.org/abs/2507.21678"}
{"aliases": [], "authors": ["Yi Kong", "Dianxi Shi", "Guoli Yang", "Zhang ke-di", "Chenlin Huang", "Xiaopeng Li", "Songchang Jin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "197dbb0e2734f30349260444fca80a8ac830a5d4000a677d6417f46509737475", "submitted_at": "2025-07-29T16:05:32Z", "updated_at": "2025-07-29T16:05:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.21953v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2507.21953", "identifiers": {"arxiv": "2507.21953"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.21953"}], "name": "MapAgent: Trajectory-Constructed Memory-Augmented Planning for Mobile Task Automation", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-07-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds reusable page memories from past mobile trajectories and retrieves them to support task planning and execution.", "topics": ["memory", "planning", "mobile-navigation"], "updated_at": "2025-07-29", "url": "https://arxiv.org/abs/2507.21953"}
{"aliases": [], "authors": ["Shuquan Lian", "Yuhang Wu", "Jia Ma", "Yifan Ding", "Zihan Song", "Bingqi Chen", "Xiawu Zheng", "Hui Li", "Rongrong Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "741cce21b25e9b29b99e9804a399695bf8a0ae89df6307d9f9bff68e3e856e85", "submitted_at": "2025-07-29T17:22:07Z", "updated_at": "2026-04-08T16:03:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.22025v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.22025", "identifiers": {"arxiv": "2507.22025"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.22025"}], "name": "UI-AGILE: Advancing GUI Agents with Effective Reinforcement Learning and Precise Inference-Time Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines grounding-oriented reward and sampling improvements with inference-time image decomposition and target selection.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2026-04-08", "url": "https://arxiv.org/abs/2507.22025"}
{"aliases": [], "authors": ["Hussein Mozannar", "Gagan Bansal", "Cheng Tan", "Adam Fourney", "Victor Dibia", "Jingya Chen", "Jack Gerrits", "Tyler Payne", "Matheus Kunzler Maldaner", "Madeleine Grunde-McLaughlin", "Eric Zhu", "Griffin Bassman", "Jacob Alber", "Peter Chang", "Ricky Loynd", "Friederike Niedtner", "Ece Kamar", "Maya Murad", "Rafah Hosn", "Saleema Amershi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c9416ede89036d640412496b5f0def218b69271c5f2449689d4e4afacb43c556", "submitted_at": "2025-07-30T03:49:14Z", "updated_at": "2025-07-30T03:49:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.22358v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2507.22358", "identifiers": {"arxiv": "2507.22358"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.22358"}], "name": "Magentic-UI: Towards Human-in-the-loop Agentic Systems", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Magentic-UI combines browser interaction, code execution, file operations, and MCP tools in an interface for human-agent collaboration. It studies collaborative planning and task execution, action guards, and persistent memory through benchmark tasks, simulated users, user studies, and safety assessments.", "topics": ["human-in-the-loop", "tool-integration", "web-navigation", "safety-security"], "updated_at": "2025-07-30", "url": "https://arxiv.org/abs/2507.22358"}
{"aliases": [], "authors": ["Haoran Sun", "Shaoning Zeng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9df4716bfc177013358648147e7562af358ef700093124408943a2e4c811107d", "submitted_at": "2025-07-23T12:45:44Z", "updated_at": "2025-07-23T12:45:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.22925v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2507.22925", "identifiers": {"arxiv": "2507.22925"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.22925"}], "name": "Hierarchical Memory for High-Efficiency Long-Term Reasoning in LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-07-23", "url": "https://arxiv.org/abs/2507.22925"}
{"aliases": [], "authors": ["Ram Mohan Rao Kadiyala", "Siddhant Gupta", "Jebish Purbey", "Giulio Martini", "Ali Shafique", "Suman Debnath", "Hamza Farooq"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "21a2172cb4776b1a0f5dec89002d913b86ba463f434c452a048b2eaeb2a2c4e6", "submitted_at": "2025-07-31T08:32:37Z", "updated_at": "2025-08-06T18:41:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.23336v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2507.23336", "identifiers": {"arxiv": "2507.23336"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2507.23336"}], "name": "DSBC : Data Science task Benchmarking with Context engineering", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-06", "url": "https://arxiv.org/abs/2507.23336"}
{"aliases": [], "authors": ["Miaosen Zhang", "Ziqiang Xu", "Jialiang Zhu", "Qi Dai", "Kai Qiu", "Yifan Yang", "Chong Luo", "Tianyi Chen", "Justin Wagle", "Tim Franklin", "Baining Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a7a32f30d2ada7f739e460429b6453d0e0e56364dda8653b0511ef2e9e6df2a", "submitted_at": "2025-07-31T17:59:09Z", "updated_at": "2025-07-31T17:59:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.23779v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2507.23779", "identifiers": {"arxiv": "2507.23779"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.23779"}], "name": "Phi-Ground Tech Report: Advancing Perception in GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Reports a GUI-grounding model family and empirical design choices covering source data, perception training, and coordinate prediction.", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-07-31", "url": "https://arxiv.org/abs/2507.23779"}
{"aliases": [], "authors": ["Aaron Xuxiang Tian", "Ruofan Zhang", "Janet Tang", "Ji Wang", "Tianyu Shi", "Jiaxin Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6edb459a166897cbf878eea707817ba2f89c32968076ec14860fe0a619c33ddb", "submitted_at": "2025-07-31T07:02:19Z", "updated_at": "2026-09-03T03:59:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.00935v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.00935", "identifiers": {"arxiv": "2508.00935"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.00935"}], "name": "Measuring Harmfulness of Computer-Using Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-07-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-03", "url": "https://arxiv.org/abs/2508.00935"}
{"aliases": [], "authors": ["Zihan Zheng", "Tianle Cui", "Taoran Wang", "Fengtao Wang", "Jiahui Pan", "Lewei He", "Qianglong Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57d89909534af90cbc10b97a992238bde3ae976ac6cc97778b4345635cef552c", "submitted_at": "2025-08-02T11:53:41Z", "updated_at": "2026-04-20T02:44:32Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2508.01330v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2508.01330", "identifiers": {"arxiv": "2508.01330"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.01330"}], "name": "NaturalGAIA: A Verifiable Benchmark and Hierarchical Framework for Long-Horizon GUI Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "NaturalGAIA builds verifiable long-horizon GUI tasks from human interaction intents while separating causal task structure from its verbal description. Its LightManus-Jarvis framework divides dynamic planning and context management from execution based on visual and structural interface information.", "topics": ["evaluation", "long-horizon", "planning"], "updated_at": "2026-04-20", "url": "https://arxiv.org/abs/2508.01330"}
{"aliases": [], "authors": ["Quan-Sheng Zeng", "Yunheng Li", "Qilong Wang", "Peng-Tao Jiang", "Zuxuan Wu", "Ming-Ming Cheng", "Qibin Hou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7bad83b404b833d44c4586995e9be135cb2c0e510e12f37e6afde5e1224cae29", "submitted_at": "2025-08-03T02:15:43Z", "updated_at": "2025-08-03T02:15:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.01548v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.01548", "identifiers": {"arxiv": "2508.01548"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.01548"}], "name": "A Glimpse to Compress: Dynamic Visual Token Pruning for Large Vision-Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-03", "url": "https://arxiv.org/abs/2508.01548"}
{"aliases": [], "authors": ["Yuhan Guo", "Cong Guo", "Aiwen Sun", "Hongliang He", "Xinyu Yang", "Yue Lu", "Yingji Zhang", "Xuntao Guo", "Dong Zhang", "Jianzhuang Liu", "Jiang Duan", "Yijia Xiao", "Liangjian Wen", "Hai-Ming Xu", "Yong Dai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7406b48ed9826d51215d955c31c3c67cd2ce9e08bf9a93a55ac4e85a546315f1", "submitted_at": "2025-08-03T17:17:52Z", "updated_at": "2026-07-05T06:17:25Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.01858v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:13+00:00", "role": "primary", "sha256": "1c8ad21928ad5e15bba474a060d225b7da000b37878a8e98dfe5ce94008e3086", "supports": ["scope", "environments"], "url": "https://raw.githubusercontent.com/Gnonymous/Web-CogReasoner/HEAD/README.md"}], "id": "arxiv:2508.01858", "identifiers": {"arxiv": "2508.01858"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.01858"}], "name": "Web-CogReasoner: Towards Multimodal Knowledge-Induced Cognitive Reasoning for Web Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-08-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Web-CogReasoner separates factual and conceptual web knowledge from procedural reasoning, using a structured curriculum to train a multimodal agent. Web-CogDataset and Web-CogBench support study of these capabilities alongside web-navigation evaluation.", "topics": ["data-generation", "planning", "evaluation", "web-navigation"], "updated_at": "2026-07-05", "url": "https://arxiv.org/abs/2508.01858"}
{"aliases": [], "authors": ["Zhihao Luo", "Wentao Yan", "Jingyu Gong", "Min Wang", "Zhizhong Zhang", "Xuhong Wang", "Yuan Xie", "Xin Tan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e892ab93511f4164f37ef79ac0473e3c1eed7d6ec11fb6c8ec82092f26c5ca68", "submitted_at": "2025-08-04T04:28:18Z", "updated_at": "2026-05-02T12:32:44Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2508.02046v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2508.02046", "identifiers": {"arxiv": "2508.02046"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.02046"}], "name": "NaviMaster: Learning a Unified Policy for GUI and Embodied Navigation Tasks", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2025-08-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "NaviMaster trains a shared policy for GUI and embodied navigation using a common visual-trajectory formulation. It combines trajectory collection, reinforcement learning and distance-aware rewards, with evaluation across both task families.", "topics": ["reinforcement-learning", "data-generation", "robustness"], "updated_at": "2026-05-02", "url": "https://arxiv.org/abs/2508.02046"}
{"aliases": [], "authors": ["Renan Souza", "Amal Gueroudji", "Stephen DeWitt", "Daniel Rosendo", "Tirthankar Ghosal", "Robert Ross", "Prasanna Balaprakash", "Rafael Ferreira da Silva"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "080c0698091725ed812c2dccb717a86907987dba3e5dd9c24f11eaa049d50eb8", "submitted_at": "2025-08-04T19:54:40Z", "updated_at": "2025-08-20T15:00:50Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.02866v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.02866", "identifiers": {"arxiv": "2508.02866"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.02866"}], "name": "PROV-AGENT: Unified Provenance for Tracking AI Agent Interactions in Agentic Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-20", "url": "https://arxiv.org/abs/2508.02866"}
{"aliases": [], "authors": ["Liujian Tang", "Shaokang Dong", "Yijia Huang", "Minqi Xiang", "Hongtao Ruan", "Bin Wang", "Shuo Li", "Zhiheng Xi", "Zhihui Cao", "Hailiang Pang", "Heng Kong", "He Yang", "Mingxu Chai", "Zhilin Gao", "Xingyu Liu", "Yingnan Fu", "Jiaming Liu", "Xuanjing Huang", "Yu-Gang Jiang", "Tao Gui", "Qi Zhang", "Kang Wang", "Yunke Zhang", "Yuran Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5175e5ed0061660fa950512300daa878eb90e5c98bdac8f80d478414b0a2931", "submitted_at": "2025-07-19T12:33:43Z", "updated_at": "2025-09-11T14:28:11Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2508.03700v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.03700", "identifiers": {"arxiv": "2508.03700"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.03700"}], "name": "MagicGUI: A Foundational Mobile GUI Agent with Scalable Data Pipeline and Reinforcement Fine-tuning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-07-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines a mobile GUI data pipeline, perception and planning capabilities, and reinforcement fine-tuning for interactive task execution.", "topics": ["mobile-navigation", "data-generation", "reinforcement-learning"], "updated_at": "2025-09-11", "url": "https://arxiv.org/abs/2508.03700"}
{"aliases": [], "authors": ["Linxin Song", "Yutong Dai", "Viraj Prabhu", "Jieyu Zhang", "Taiwei Shi", "Li Li", "Junnan Li", "Silvio Savarese", "Zeyuan Chen", "Jieyu Zhao", "Ran Xu", "Caiming Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6e4272a69c1a93f22bd35032547ffb6021a79c97d179a1075e14aee69dea83a", "submitted_at": "2025-08-05T21:33:36Z", "updated_at": "2026-02-20T10:14:08Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.03923v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.03923", "identifiers": {"arxiv": "2508.03923"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.03923"}], "name": "CoAct-1: Computer-using Multi-Agent System with Coding Actions", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "CoAct-1 delegates subtasks between a GUI operator and a programmer that executes Python or Bash. Its OSWorld evaluation studies a hybrid action space in which scripts handle suitable file and data operations while visual interaction handles other steps.", "topics": ["tool-integration", "automation", "efficiency", "planning"], "updated_at": "2026-02-20", "url": "https://arxiv.org/abs/2508.03923"}
{"aliases": [], "authors": ["Zheyuan Zhou", "Jiayi Han", "Liang Du", "Naiyu Fang", "Lemiao Qiu", "Shuyou Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4f8a21e2b0d2c1a8ec55e75ed5c4b014e7b1f47df3f8a3e09db633a9035e099", "submitted_at": "2025-08-06T01:30:56Z", "updated_at": "2025-08-06T01:30:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04002v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.04002", "identifiers": {"arxiv": "2508.04002"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.04002"}], "name": "CAD-Judge: Toward Efficient Morphological Grading and Verification for Text-to-CAD Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-06", "url": "https://arxiv.org/abs/2508.04002"}
{"aliases": [], "authors": ["Yurun Chen", "Xavier Hu", "Yuhan Liu", "Keting Yin", "Juncheng Li", "Zhuosheng Zhang", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0ff68ce8b01bd31135b3639c0613977539d5546983ab35b934be333ad66dd32", "submitted_at": "2025-08-06T01:49:32Z", "updated_at": "2025-08-06T01:49:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04010v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.04010", "identifiers": {"arxiv": "2508.04010"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.04010"}], "name": "HarmonyGuard: Toward Safety and Utility in Web Agents via Adaptive Policy Enhancement and Dual-Objective Optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-06", "url": "https://arxiv.org/abs/2508.04010"}
{"aliases": [], "authors": ["Chao Hao", "Shuai Wang", "Kaiwen Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "499a462f9f14b8afdd367bb7cf7a7ffb1b2e166c212fb5c36686d3145f08ab5f", "submitted_at": "2025-08-06T02:38:02Z", "updated_at": "2025-08-06T02:38:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04025v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.04025", "identifiers": {"arxiv": "2508.04025"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04025"}], "name": "Uncertainty-Aware GUI Agent: Adaptive Perception through Component Recommendation and Human-in-the-Loop Refinement", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "RecAgent narrows attention to recommended UI components and requests user feedback when task intent or decisions remain uncertain. It introduces ComplexAction to evaluate specified single-step GUI actions in difficult interface scenarios.", "topics": ["mobile-navigation", "human-in-the-loop", "screen-understanding", "evaluation"], "updated_at": "2025-08-06", "url": "https://arxiv.org/abs/2508.04025"}
{"aliases": [], "authors": ["Shunyu Liu", "Minghao Liu", "Huichi Zhou", "Zhenyu Cui", "Yang Zhou", "Yuhao Zhou", "Jialiang Gao", "Heng Zhou", "Yunhao Yang", "Wendong Fan", "puzhen zhang", "Ge Zhang", "Jiajun Shi", "Weihao Xuan", "Jiaxing Huang", "Shuang Luo", "Fang Wu", "Heli Qi", "Qingcheng Zeng", "Junjie Wang", "Aosong Feng", "Jindi Lv", "Sicong Jiang", "Ziqi Ren", "Wangchunshu Zhou", "Zhenfei Yin", "Wenlong Zhang", "Guohao Li", "Wenhao Yu", "Lei Ma", "Lei Bai", "Qunshu Lin", "Mingli Song", "Dacheng Tao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2746f8bea42f6153e7ff009f28d334c662a3d7d6826d0d8b94a1bbe97b07e325", "submitted_at": "2025-08-06T02:38:18Z", "updated_at": "2026-02-27T14:53:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.04026v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2508.04026", "identifiers": {"arxiv": "2508.04026"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.04026"}], "name": "VeriWeb: Verifiable Long-Chain Web Benchmark for Agentic Information-Seeking", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-27", "url": "https://arxiv.org/abs/2508.04026"}
{"aliases": [], "authors": ["Yuhao Cheng", "Liang Tang", "Shuxian Li", "Yukang Huo", "Tiaonan Duan", "Kaer Huang", "Yanzhe Jing", "Yiqiang Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d54ad74695a60cb040d88fcb85549912d7d4ecf6ca2ed78e636d05f8f17d1795", "submitted_at": "2025-08-06T02:57:22Z", "updated_at": "2026-01-22T05:22:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.04037v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.04037", "identifiers": {"arxiv": "2508.04037"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04037"}], "name": "Evolving in Tasks: Empowering the Multi-modality Large Language Model as the Computer Use Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "The Self-Evolution Agent combines automatic generation of verifiable computer-use trajectories with step-wise reinforcement learning for long-horizon training. It also describes a model-enhancement procedure that integrates grounding and planning capabilities without an additional training stage.", "topics": ["data-generation", "reinforcement-learning", "grounding", "planning"], "updated_at": "2026-01-22", "url": "https://arxiv.org/abs/2508.04037"}
{"aliases": [], "authors": ["Weitai Kang", "Bin Lei", "Gaowen Liu", "Caiwen Ding", "Yan Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87b5336c1850c15228880e0625252331b15b79d1cf1e3929abbc01ead70d5f25", "submitted_at": "2025-08-06T12:35:24Z", "updated_at": "2025-08-06T12:35:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04389v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:37Z", "role": "primary", "sha256": "afdfecfc117506609a5bff392254299d88cc84c971469efe0601cbcb02660ecf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2508.04389v1"}], "id": "arxiv:2508.04389", "identifiers": {"arxiv": "2508.04389"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04389"}], "name": "GuirlVG: Incentivize GUI Visual Grounding via Empirical Exploration on Reinforcement Learning", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GuirlVG, an empirical study of reinforcement fine-tuning for GUI visual grounding. It analyzes training components and adds an adaptive KL-based mechanism to limit reward over-optimization.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-08-06", "url": "https://arxiv.org/abs/2508.04389"}
{"aliases": [], "authors": ["Thassilo M. Schiepanski", "Nicholas Piël"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "789cfb56a83974c38e6f5e359dac5b6b3075cb70ac8afc178e36a901ce742041", "submitted_at": "2025-08-06T12:56:54Z", "updated_at": "2026-08-31T14:39:02Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2508.04412v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.04412", "identifiers": {"arxiv": "2508.04412"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04412"}], "name": "Beyond Pixels: Exploring DOM Downsampling for LLM-Based Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "D2Snap reduces serialized DOM observations while preserving actionable elements and features that distinguish their uses. The study compares reduced DOM inputs with screenshot-based observations for a web agent on tasks sampled from Online-Mind2Web.", "topics": ["web-navigation", "efficiency", "screen-understanding"], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2508.04412"}
{"aliases": [], "authors": ["Xueyu Hu", "Tao Xiong", "Biao Yi", "Zishu Wei", "Ruixuan Xiao", "Yurun Chen", "Jiasheng Ye", "Meiling Tao", "Xiangxin Zhou", "Ziyu Zhao", "Yuhuai Li", "Shengze Xu", "Shenzhi Wang", "Xinchen Xu", "Shuofei Qiao", "Zhaokai Wang", "Kun Kuang", "Tieyong Zeng", "Liang Wang", "Jiwei Li", "Yuchen Eleanor Jiang", "Wangchunshu Zhou", "Guoyin Wang", "Keting Yin", "Zhou Zhao", "Hongxia Yang", "Fan Wu", "Shengyu Zhang", "Fei Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f14fdd895e52277a3695eeb83e5095472051d139aec63f22f9c63d344c051dad", "submitted_at": "2025-08-06T14:33:45Z", "updated_at": "2025-08-06T14:33:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04482v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2508.04482", "identifiers": {"arxiv": "2508.04482"}, "kind": "survey", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04482"}], "name": "OS Agents: A Survey on MLLM-based Agents for General Computing Devices Use", "notes": "", "published_at": "2025-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Surveys agents that operate computers and phones through operating-system interfaces, covering components, capabilities, foundation models, frameworks, and benchmarks.", "topics": ["survey"], "updated_at": "2025-08-06", "url": "https://arxiv.org/abs/2508.04482"}
{"aliases": [], "authors": ["Zeyi Sun", "Ziyu Liu", "Yuhang Zang", "Yuhang Cao", "Xiaoyi Dong", "Tong Wu", "Dahua Lin", "Jiaqi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "268c38e6f11f11771a1251fc522112ff052557b5753dff6cadf2ba49c0ec04b2", "submitted_at": "2025-08-06T17:58:46Z", "updated_at": "2025-08-12T15:11:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.04700v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.04700", "identifiers": {"arxiv": "2508.04700"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04700"}], "name": "SEAgent: Self-Evolving Computer Use Agent with Autonomous Learning from Experience", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "SEAgent explores unfamiliar software using a world-state assessor and an automatically generated task curriculum. It learns from failed and successful interactions through adversarial imitation and GRPO, then combines specialist experience into a generalist agent evaluated on five OSWorld software environments.", "topics": ["reinforcement-learning", "data-generation", "world-models", "imitation-learning"], "updated_at": "2025-08-12", "url": "https://arxiv.org/abs/2508.04700"}
{"aliases": [], "authors": ["Jiarun Liu", "Chunhong Zhang", "Zheng Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d574d4157b775639c08ed12813892584b011dd3c0271ca2586f9d6a61d00c14", "submitted_at": "2025-08-07T07:05:22Z", "updated_at": "2025-08-07T07:05:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.05081v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.05081", "identifiers": {"arxiv": "2508.05081"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.05081"}], "name": "Cognitive Duality for Adaptive Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-08-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CogniWeb, which switches between fast reactive processing and slower deliberative reasoning according to task complexity. WebArena experiments examine this combination of learned behaviors and online planning.", "topics": ["planning", "efficiency", "web-navigation"], "updated_at": "2025-08-07", "url": "https://arxiv.org/abs/2508.05081"}
{"aliases": [], "authors": ["Kaiwen Yan", "Yuhang Chang", "Zirui Guo", "Yaling Mou", "Jiang Ming", "Jingwei Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "be94532821a2d7fdbee82066bdacf8bec8adf7d874e7deed9eda347105138755", "submitted_at": "2025-08-07T09:28:43Z", "updated_at": "2025-08-07T09:28:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.05193v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.05193", "identifiers": {"arxiv": "2508.05193"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.05193"}], "name": "STEPWISE-CODEX-Bench: Evaluating Complex Multi-Function Comprehension and Fine-Grained Execution Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-07", "url": "https://arxiv.org/abs/2508.05193"}
{"aliases": [], "authors": ["Sahar Salimpour", "Lei Fu", "Kajetan Rachwał", "Pascal Bertrand", "Kevin O'Sullivan", "Robert Jakob", "Farhad Keramat", "Leonardo Militano", "Giovanni Toffetti", "Harry Edelman", "Jorge Peña Queralta"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "62076930f3b9024f3de7f8e1e3230d32cef98f16d00001ae3f3eefe58b20615f", "submitted_at": "2025-08-07T11:48:03Z", "updated_at": "2025-11-12T19:04:19Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2508.05294v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.05294", "identifiers": {"arxiv": "2508.05294"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.05294"}], "name": "Towards Embodied Agentic AI: Review and Classification of LLM- and VLM-Driven Robot Autonomy and Interaction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-12", "url": "https://arxiv.org/abs/2508.05294"}
{"aliases": [], "authors": ["Yong Du", "Yuchen Yan", "Fei Tang", "Zhengxi Lu", "Chang Zong", "Weiming Lu", "Shengpei Jiang", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c7deba3ec19a878cfd402c5dac61e4dca284d478429bc9204264448062299e99", "submitted_at": "2025-08-07T17:54:27Z", "updated_at": "2025-11-13T12:54:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.05615v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.05615", "identifiers": {"arxiv": "2508.05615"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.05615"}], "name": "Test-Time Reinforcement Learning for GUI Grounding via Region Consistency", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-RC aggregates multiple grounding predictions into spatial voting grids to choose a consensus region. GUI-RCPO then turns agreement with that region into a reward for test-time policy updates on unlabeled GUI screenshots.", "topics": ["grounding", "reinforcement-learning", "training-free"], "updated_at": "2025-11-13", "url": "https://arxiv.org/abs/2508.05615"}
{"aliases": [], "authors": ["Yuhang Liu", "Zeyu Liu", "Shuanghe Zhu", "Pengxiang Li", "Congkai Xie", "Jiasheng Wang", "Xavier Hu", "Xiaotian Han", "Jianbo Yuan", "Xinyao Wang", "Shengyu Zhang", "Hongxia Yang", "Fei Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a69d2bdb4d01e6e5c9f08da33281c8247f59af52f3bc08dfed6b7e24da88e194", "submitted_at": "2025-08-07T17:49:56Z", "updated_at": "2025-12-08T10:49:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.05731v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.05731", "identifiers": {"arxiv": "2508.05731"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.05731"}], "name": "InfiGUI-G1: Advancing GUI Grounding with Adaptive Exploration Policy Optimization", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses multiple candidate answers and adaptive exploration rewards to improve spatial and semantic alignment in GUI grounding.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-12-08", "url": "https://arxiv.org/abs/2508.05731"}
{"aliases": [], "authors": ["Xinyu Geng", "Peng Xia", "Zhen Zhang", "Xinyu Wang", "Qiuchen Wang", "Ruixue Ding", "Chenxi Wang", "Jialong Wu", "Yida Zhao", "Kuan Li", "Yong Jiang", "Pengjun Xie", "Fei Huang", "Jingren Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bdbd38ef30e655bfd562f97007fc8eb4b4db57596489f5292d93af64552d6d6", "submitted_at": "2025-08-07T18:03:50Z", "updated_at": "2025-09-01T03:21:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.05748v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.05748", "identifiers": {"arxiv": "2508.05748"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.05748"}], "name": "WebWatcher: Breaking New Frontier of Vision-Language Deep Research Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-01", "url": "https://arxiv.org/abs/2508.05748"}
{"aliases": [], "authors": ["Keyu Li", "Mohan Jiang", "Dayuan Fu", "Yunze Wu", "Xiangkun Hu", "Dequan Wang", "Pengfei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e471175e25a5f0b6f4d3643a215dd43702c0defa996fc41a75bc50581c33730", "submitted_at": "2025-08-09T12:15:08Z", "updated_at": "2025-08-09T12:15:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.06960v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2508.06960", "identifiers": {"arxiv": "2508.06960"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.06960"}], "name": "DatasetResearch: Benchmarking Agent Systems for Demand-Driven Dataset Discovery", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-09", "url": "https://arxiv.org/abs/2508.06960"}
{"aliases": [], "authors": ["Yanshu Li", "Jianjiang Yang", "Zhennan Shen", "Ligong Han", "Haoyan Xu", "Ruixiang Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c7d30174ec1665009b475090d81b17efdfffe8693c9f4f67b400a518abf92be", "submitted_at": "2025-08-11T11:41:51Z", "updated_at": "2025-12-09T06:05:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.07871v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.07871", "identifiers": {"arxiv": "2508.07871"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.07871"}], "name": "CATP: Contextually Adaptive Token Pruning for Efficient and Enhanced Multimodal In-Context Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-09", "url": "https://arxiv.org/abs/2508.07871"}
{"aliases": [], "authors": ["Ryan Wong", "Jiawei Wang", "Junjie Zhao", "Li Chen", "Yan Gao", "Long Zhang", "Xuan Zhou", "Zuo Wang", "Kai Xiang", "Ge Zhang", "Wenhao Huang", "Yang Wang", "Ke Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c55f58d674035913cb6b7551c0663518a78f4045e00135bfd2d3ed199939fd9", "submitted_at": "2025-08-11T14:03:09Z", "updated_at": "2025-08-28T09:31:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.07999v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.07999", "identifiers": {"arxiv": "2508.07999"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.07999"}], "name": "WideSearch: Benchmarking Agentic Broad Info-Seeking", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-28", "url": "https://arxiv.org/abs/2508.07999"}
{"aliases": [], "authors": ["Sining Lu", "Guan Chen", "Nam Anh Dinh", "Itai Lang", "Ari Holtzman", "Rana Hanocka"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c1bf2467176caf5be7ddc6ec42c4872ffd880ce65f0bf1da0d4b2f7e3f23447", "submitted_at": "2025-08-11T17:48:02Z", "updated_at": "2025-08-11T17:48:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.08228v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.08228", "identifiers": {"arxiv": "2508.08228"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.08228"}], "name": "LL3M: Large Language 3D Modelers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-11", "url": "https://arxiv.org/abs/2508.08228"}
{"aliases": [], "authors": ["Sizhe Yuen", "Francisco Gomez Medina", "Ting Su", "Yali Du", "Adam J. Sobey"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a7aa5ee50fa6890749471907aaf5c96d785977bbdaadb915433c3e44734b63c", "submitted_at": "2025-08-12T15:05:00Z", "updated_at": "2026-01-12T11:46:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.08997v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2508.08997", "identifiers": {"arxiv": "2508.08997"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.08997"}], "name": "Intrinsic Memory Agents: Heterogeneous Multi-Agent LLM Systems through Structured Contextual Memory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2508.08997"}
{"aliases": [], "authors": ["Zeyu Huang", "Juyuan Wang", "Longfeng Chen", "Boyi Xiao", "Leng Cai", "Yawen Zeng", "Jin Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4efa329c3b106219c9b25f675be0c561743085ac999d2964458a5f22ff6e62f", "submitted_at": "2025-08-12T16:18:30Z", "updated_at": "2025-08-14T16:36:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.09057v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2508.09057", "identifiers": {"arxiv": "2508.09057"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.09057"}], "name": "MVISU-Bench: Benchmarking Mobile Agents for Real-World Tasks by Multi-App, Vague, Interactive, Single-App and Unethical Instructions", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-08-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "MVISU-Bench evaluates bilingual mobile tasks involving cross-app work, vague requests, user interaction and unethical instructions alongside ordinary single-app tasks. Its Aider prompting module is designed to clarify intent and mitigate risky execution within agent workflows.", "topics": ["evaluation", "mobile-navigation", "human-in-the-loop", "safety-security"], "updated_at": "2025-08-14", "url": "https://arxiv.org/abs/2508.09057"}
{"aliases": [], "authors": ["Xinyuan Wang", "Bowen Wang", "Dunjie Lu", "Junlin Yang", "Tianbao Xie", "Junli Wang", "Jiaqi Deng", "Xiaole Guo", "Yiheng Xu", "Chen Henry Wu", "Zhennan Shen", "Zhuokai Li", "Ryan Li", "Xiaochuan Li", "Junda Chen", "Boyuan Zheng", "Peihang Li", "Fangyu Lei", "Ruisheng Cao", "Yeqiao Fu", "Dongchan Shin", "Martin Shin", "Jiarui Hu", "Yuyan Wang", "Jixuan Chen", "Yuxiao Ye", "Danyang Zhang", "Dikang Du", "Hao Hu", "Huarong Chen", "Zaida Zhou", "Haotian Yao", "Ziwei Chen", "Qizheng Gu", "Yipu Wang", "Heng Wang", "Diyi Yang", "Victor Zhong", "Flood Sung", "Y. Charles", "Zhilin Yang", "Tao Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40e46f27128d066358fcee6279b26bb2b708dacaf1f7b5794b755520fe892243", "submitted_at": "2025-08-12T17:52:32Z", "updated_at": "2025-10-04T17:38:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.09123v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.09123", "identifiers": {"arxiv": "2508.09123"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.09123"}], "name": "OpenCUA: Open Foundations for Computer-Use Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2025-08-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents demonstration collection infrastructure, the AgentNet dataset, and a training pipeline for open computer-use models.", "topics": ["data-generation", "planning"], "updated_at": "2025-10-04", "url": "https://arxiv.org/abs/2508.09123"}
{"aliases": [], "authors": ["Weixuan Wang", "Dongge Han", "Daniel Madrigal Diaz", "Jin Xu", "Victor Rühle", "Saravan Rajmohan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68fc1b1ebc8a06252ff59bbb5be0c84e04e249297268750093e59ff9757d1449", "submitted_at": "2025-08-12T17:53:03Z", "updated_at": "2025-08-12T17:53:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.09124v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.09124", "identifiers": {"arxiv": "2508.09124"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.09124"}], "name": "OdysseyBench: Evaluating LLM Agents on Long-Horizon Complex Office Application Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-12", "url": "https://arxiv.org/abs/2508.09124"}
{"aliases": [], "authors": ["Fengxian Ji", "Jingpu Yang", "Zirui Song", "Yuanxi Wang", "Zhexuan Cui", "Yuke Li", "Qian Jiang", "Miao Fang", "Xiuying Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8c91a1f85602760c706e56426c0d2d870bbc101ca500aa3c7c391938c3f1676", "submitted_at": "2025-08-12T15:12:42Z", "updated_at": "2025-08-12T15:12:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.09241v1"}}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.09241", "identifiers": {"arxiv": "2508.09241"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.09241"}], "name": "FineState-Bench: A Comprehensive Benchmark for Fine-Grained State Control in GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "FineState-Bench evaluates fine-grained GUI perception and control across desktop, web, and mobile interfaces. Its Visual Diagnostic Assistant separates perception and localization contributions in controlled interaction experiments.", "topics": ["evaluation", "grounding", "observability", "screen-understanding"], "updated_at": "2025-08-12", "url": "https://arxiv.org/abs/2508.09241"}
{"aliases": [], "authors": ["Chengtao Lv", "Bilang Zhang", "Yang Yong", "Ruihao Gong", "Yushi Huang", "Shiqiao Gu", "Jiajun Wu", "Yumeng Shi", "Jinyang Guo", "Wenya Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "16ce6a3892dc02d73eed4e8d7032ee732b28e3b1b983790806454f1c57d1f5a9", "submitted_at": "2025-08-13T17:54:49Z", "updated_at": "2025-11-17T13:22:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.09981v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.09981", "identifiers": {"arxiv": "2508.09981"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.09981"}], "name": "LLMC+: Benchmarking Vision-Language Model Compression with a Plug-and-play Toolkit", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-17", "url": "https://arxiv.org/abs/2508.09981"}
{"aliases": [], "authors": ["Zhangxuan Gu", "Zhengwen Zeng", "Zhenyu Xu", "Xingran Zhou", "Shuheng Shen", "Yunfei Liu", "Beitong Zhou", "Changhua Meng", "Tianyu Xia", "Weizhi Chen", "Yue Wen", "Jingya Dou", "Fei Tang", "Jinzhen Lin", "Yulin Liu", "Zhenlin Guo", "Yichen Gong", "Heng Jia", "Changlong Gao", "Yuan Guo", "Yong Deng", "Zhenyu Guo", "Liang Chen", "Weiqiang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db88c4ba15c48210f0b618a8b5d2b9c02184f2e4375d45adf22b9a0a395abc45", "submitted_at": "2025-08-14T16:58:07Z", "updated_at": "2025-08-15T14:49:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.10833v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2508.10833", "identifiers": {"arxiv": "2508.10833"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.10833"}], "name": "UI-Venus Technical Report: Building High-performance UI Agents with RFT", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-08-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies screenshot-based GUI grounding and navigation using curated interaction data, reward design, and reinforcement fine-tuning.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-08-15", "url": "https://arxiv.org/abs/2508.10833"}
{"aliases": [], "authors": ["Songqin Nong", "Xiaoxuan Tang", "Jingxuan Xu", "Sheng Zhou", "Jianfeng Chen", "Tao Jiang", "Wenhao Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea849d8e7640d7a40d582431233418cf1cbd29585ca6c78638919b8832265445", "submitted_at": "2025-08-15T09:55:02Z", "updated_at": "2026-03-13T11:10:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.11360v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2508.11360", "identifiers": {"arxiv": "2508.11360"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.11360"}], "name": "CRAFT-GUI: Curriculum-Reinforced Agent For GUI Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "CRAFT-GUI organizes GUI trajectories by difficulty for curriculum-based GRPO training. Its reward combines rule-based signals with model judgments, with evaluation on Android Control and an internal online benchmark.", "topics": ["reinforcement-learning", "reward-models", "mobile-navigation"], "updated_at": "2026-03-13", "url": "https://arxiv.org/abs/2508.11360"}
{"aliases": [], "authors": ["Ralph Peeters", "Aaron Steiner", "Luca Schwarz", "Julian Yuya Caspary", "Christian Bizer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0307d242a90f0650b5d5fdfbaf9653842d3ce1461d2653b0b1ba6b21c31ee2dc", "submitted_at": "2025-08-18T15:41:22Z", "updated_at": "2026-04-30T09:17:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.13024v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.13024", "identifiers": {"arxiv": "2508.13024"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.13024"}], "name": "WebMall -- A Multi-Shop Benchmark for Evaluating Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebMall provides four simulated online shops with heterogeneous product data for comparison shopping, related-product searches, and checkout tasks. It evaluates web agents with different observation formats and memory configurations in a fixed offline environment.", "topics": ["evaluation", "web-navigation", "sandbox"], "updated_at": "2026-04-30", "url": "https://arxiv.org/abs/2508.13024"}
{"aliases": [], "authors": ["Weizhen Li", "Jianbo Lin", "Zhuosong Jiang", "Jingyi Cao", "Xinpeng Liu", "Jiayu Zhang", "Zhenqiang Huang", "Qianben Chen", "Weichen Sun", "Qiexiang Wang", "Hongxuan Lu", "Tianrui Qin", "Chenghao Zhu", "Yi Yao", "Shuying Fan", "Xiaowan Li", "Tiannan Wang", "Pai Liu", "King Zhu", "He Zhu", "Dingfeng Shi", "Piaohong Wang", "Yeyi Guan", "Xiangru Tang", "Minghao Liu", "Yuchen Eleanor Jiang", "Jian Yang", "Jiaheng Liu", "Ge Zhang", "Wangchunshu Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de15b704965fa2daf2e4940e3bf8d3312b38e17c18434a9cd3a0c843f85724e8", "submitted_at": "2025-08-06T17:01:02Z", "updated_at": "2025-08-06T17:01:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.13167v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2508.13167", "identifiers": {"arxiv": "2508.13167"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.13167"}], "name": "Chain-of-Agents: End-to-End Agent Foundation Models via Multi-Agent Distillation and Agentic RL", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-06", "url": "https://arxiv.org/abs/2508.13167"}
{"aliases": [], "authors": ["Shilong Li", "Xingyuan Bu", "Wenjie Wang", "Jiaheng Liu", "Jun Dong", "Haoyang He", "Hao Lu", "Haozhe Zhang", "Chenchen Jing", "Zhen Li", "Chuanhao Li", "Jiayi Tian", "Chenchen Zhang", "Tianhao Peng", "Yancheng He", "Jihao Gu", "Hui Huang", "Donghao Zhou", "Yuanxing Zhang", "Jian Yang", "Ge Zhang", "Wenhao Huang", "Zhaoxiang Zhang", "Qiangpeng Yang", "Shilei Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1576451eea518cb7de11d96556874ffecd4e7c3e5025b2823b799a9ebb0d78b8", "submitted_at": "2025-08-14T13:46:47Z", "updated_at": "2026-08-31T07:06:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.13186v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2508.13186", "identifiers": {"arxiv": "2508.13186"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.13186"}], "name": "MM-BrowseComp: A Comprehensive Benchmark for Multimodal Browsing Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2508.13186"}
{"aliases": [], "authors": ["Jikai Chen", "Long Chen", "Dong Wang", "Qinglin Su", "Zhixuan Chu", "Bingguang Hao", "Leilei Gan", "Chenyi Zhuang", "Jinjie Gu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfb50e10184cded6c18d4c49401145bc7cfa00b0ba21bce964abe8f459af5626", "submitted_at": "2025-08-19T08:47:44Z", "updated_at": "2026-01-20T06:39:19Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.13634v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.13634", "identifiers": {"arxiv": "2508.13634"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.13634"}], "name": "V2P: Visual Attention Calibration for GUI Grounding via Background Suppression and Center Peaking", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "V2P trains GUI grounding by suppressing attention to background regions and emphasizing the center of the target element. It represents target importance with size-dependent Gaussian heatmaps and evaluates the contribution of each component on ScreenSpot-v2 and ScreenSpot-Pro.", "topics": ["grounding", "screen-understanding"], "updated_at": "2026-01-20", "url": "https://arxiv.org/abs/2508.13634"}
{"aliases": [], "authors": ["Hanyu Lai", "Xiao Liu", "Yanxiao Zhao", "Han Xu", "Hanchen Zhang", "Bohao Jing", "Yanyu Ren", "Shuntian Yao", "Yuxiao Dong", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "383d8e80619d58e72ebb9c49229cd9522a0b3586b4bfa8f00b29b1834adca5b9", "submitted_at": "2025-08-19T17:59:45Z", "updated_at": "2025-10-21T05:18:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.14040v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.14040", "identifiers": {"arxiv": "2508.14040"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.14040"}], "name": "ComputerRL: Scaling End-to-End Online Reinforcement Learning for Computer Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines API and GUI actions with distributed desktop reinforcement learning and alternating supervised training to address training instability.", "topics": ["reinforcement-learning", "tool-integration"], "updated_at": "2025-10-21", "url": "https://arxiv.org/abs/2508.14040"}
{"aliases": [], "authors": ["Yutong Bian", "Xianhao Lin", "Yupeng Xie", "Tianyang Liu", "Mingchen Zhuge", "Siyuan Lu", "Haoming Tang", "Jinlin Wang", "Jiayi Zhang", "Jiaqi Chen", "Xiangru Tang", "Yongxin Ni", "Sirui Hong", "Chenglin Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5034c06e63607a1279668964e91c14ed762a95b6e3953337ef8b8b63c4fd57d", "submitted_at": "2025-08-17T07:31:11Z", "updated_at": "2025-08-17T07:31:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.14104v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2508.14104", "identifiers": {"arxiv": "2508.14104"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.14104"}], "name": "You Don't Know Until You Click:Automated GUI Testing for Production-Ready Software Evaluation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "RealDevWorld evaluates generated software repositories with 194 development tasks and AppEvalPilot, an agent judge that operates the resulting GUIs. The evaluation examines functional behavior, visual fidelity, and runtime responses to produce task-specific diagnostic feedback.", "topics": ["evaluation", "automation", "observability"], "updated_at": "2025-08-17", "url": "https://arxiv.org/abs/2508.14104"}
{"aliases": [], "authors": ["Jiabo Ye", "Xi Zhang", "Haiyang Xu", "Haowei Liu", "Junyang Wang", "Zhaoqing Zhu", "Ziwei Zheng", "Feiyu Gao", "Junjie Cao", "Zhengxi Lu", "Jitong Liao", "Qi Zheng", "Fei Huang", "Jingren Zhou", "Ming Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "629421d60bb8beeba3f087854ebaddc5461f9af0e5fb0bc572d0c20394ba9dcd", "submitted_at": "2025-08-21T00:39:12Z", "updated_at": "2025-09-01T05:38:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.15144v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2508.15144", "identifiers": {"arxiv": "2508.15144"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.15144"}], "name": "Mobile-Agent-v3: Fundamental Agents for GUI Automation", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-08-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI-Owl models and Mobile-Agent-v3 orchestration, using environment-generated trajectories to support GUI capabilities.", "topics": ["data-generation", "planning", "mobile-navigation"], "updated_at": "2025-09-01", "url": "https://arxiv.org/abs/2508.15144"}
{"aliases": [], "authors": ["Minghao Li", "Ying Zeng", "Zhihao Cheng", "Cong Ma", "Kai Jia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e92a999dd0332536c97d986cf3bae1dbbc52cbb1f9acf6bb84e6a5a0d36d9b8", "submitted_at": "2025-08-14T03:33:43Z", "updated_at": "2025-08-14T03:33:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.15804v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.15804", "identifiers": {"arxiv": "2508.15804"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.15804"}], "name": "ReportBench: Evaluating Deep Research Agents via Academic Survey Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-14", "url": "https://arxiv.org/abs/2508.15804"}
{"aliases": [], "authors": ["Xianren Zhang", "Shreyas Prasad", "Di Wang", "Qiuhai Zeng", "Suhang Wang", "Wenbo Yan", "Mat Hans"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6c8336a707723227a9d651a3e9347e45907f28ed7421b577761ed57f7d70593d", "submitted_at": "2025-08-18T21:58:43Z", "updated_at": "2026-04-21T00:38:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.15832v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.15832", "identifiers": {"arxiv": "2508.15832"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.15832"}], "name": "A Functionality-Grounded Benchmark for Evaluating Web Agents in E-commerce Domains", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Amazon-Bench generates functionality-specific web tasks from page content and interactive controls, covering operations beyond product search. Its evaluator checks task completion and unintended changes to account state, such as address or purchase settings.", "topics": ["evaluation", "safety-security", "data-generation", "web-navigation"], "updated_at": "2026-04-21", "url": "https://arxiv.org/abs/2508.15832"}
{"aliases": [], "authors": ["Huichi Zhou", "Yihang Chen", "Siyuan Guo", "Xue Yan", "Kin Hei Lee", "Zihan Wang", "Ka Yiu Lee", "Guchun Zhang", "Kun Shao", "Linyi Yang", "Jun Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "902f573aeeb78f4a0bcd4b5c16db103de4673cadd752caa27d5473aa6c401410", "submitted_at": "2025-08-22T07:25:30Z", "updated_at": "2025-08-25T13:32:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.16153v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.16153", "identifiers": {"arxiv": "2508.16153"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.16153"}], "name": "Memento: Fine-tuning LLM Agents without Fine-tuning LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-25", "url": "https://arxiv.org/abs/2508.16153"}
{"aliases": [], "authors": ["Tanvir Bhathal", "Asanshay Gupta"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67da98d98fec647de9b482fabd8c7861063c0fdb91f1ec303fe8df759ee0f69d", "submitted_at": "2025-08-23T11:02:59Z", "updated_at": "2025-08-23T11:02:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.16987v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.16987", "identifiers": {"arxiv": "2508.16987"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.16987"}], "name": "WebSight: A Vision-First Architecture for Robust Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebSight uses a vision-language model trained on web-interface examples within a modular agent with planning, action, verification, and episodic memory components. It evaluates element clicking on Showdown Clicks and full browsing tasks on WebVoyager using visual observations.", "topics": ["web-navigation", "grounding", "planning", "memory"], "updated_at": "2025-08-23", "url": "https://arxiv.org/abs/2508.16987"}
{"aliases": [], "authors": ["Aaryaman Kartha", "Ahmed Masry", "Mohammed Saidul Islam", "Thinh Lang", "Shadikur Rahman", "Ridwan Mahbub", "Mizanur Rahman", "Mahir Ahmed", "Md Rizwan Parvez", "Enamul Hoque", "Shafiq Joty"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "def26d823e71c48d560131030876476528a09060c93d607f029819accc18b0ec", "submitted_at": "2025-08-24T15:11:44Z", "updated_at": "2025-08-24T15:11:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.17398v1"}, "publication": {"authors": ["Aaryaman Kartha", "Ahmed Masry", "Mohammed Saidul Islam", "Thinh Lang", "Shadikur Rahman", "Ridwan Mahbub", "Mizanur Rahman", "Mahir Ahmed", "Md. Rizwan Parvez", "Enamul Hoque", "Shafiq Joty"], "date_basis": "Explicit publisher metadata: 2026/3; precision=month.", "date_parts": [2026, 3], "doi": "10.18653/v1/2026.findings-eacl.177", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2026.findings-eacl.177/"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "This work already exists as metadata_verified arxiv:2508.17398. Its retained Atom body was unavailable locally, so identity is based on the exact title and full author list (minor Md. punctuation variant) in the existing record and publisher source. No arXiv API refresh or second work is proposed. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "3385-3407", "pdf_url": "https://aclanthology.org/2026.findings-eacl.177.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "DashboardQA: Benchmarking Multimodal Agents for Question Answering on Interactive Dashboards", "type": "conference-paper", "url": "https://aclanthology.org/2026.findings-eacl.177/", "venue": "Findings of the Association for Computational Linguistics: EACL 2026", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "a6c9884ad737559bdbf266858be959e6d641ade412d6cdd9db2ac6f5fa4d0a4f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2026.findings-eacl.177/"}], "id": "arxiv:2508.17398", "identifiers": {"arxiv": "2508.17398"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.17398"}, {"rel": "publication", "url": "https://aclanthology.org/2026.findings-eacl.177/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2026.findings-eacl.177.pdf"}], "name": "DashboardQA: Benchmarking Multimodal Agents for Question Answering on Interactive Dashboards", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-24", "url": "https://arxiv.org/abs/2508.17398"}
{"aliases": [], "authors": ["Xin Wang", "Zhiyao Cui", "Hao Li", "Ya Zeng", "Chenxu Wang", "Ruiqi Song", "Yihang Chen", "Kun Shao", "Qiaosheng Zhang", "Jinzhuo Liu", "Siyue Ren", "Shuyue Hu", "Zhen Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57d5fc98a2eb2232f237d71b34060df0acef89cfeaa4bc247580a4dda3223e0b", "submitted_at": "2025-08-25T13:57:02Z", "updated_at": "2025-08-25T13:57:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.18040v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:18:36.598801+00:00", "role": "primary", "sha256": "3c3474b771af7bdd500e6aaa91b83d39ec0d455e59bb83df906e7a94700ad2b7", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2508.18040v1"}], "id": "arxiv:2508.18040", "identifiers": {"arxiv": "2508.18040"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:10.946679+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.18040"}], "name": "PerPilot: Personalizing VLM-based Mobile Agents via Memory and Exploration", "notes": "Personalization depends on the validity of stored or inferred user information and retrieval access. This is instruction completion with interaction, not demonstrated permanent model-weight personalization. No specific operating system or reproduced result is asserted. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (The Framework of PerPilot; Personalization Perception; Personalization Completion; PerInstruct Data Construction); not an entire-paper reading or reproduction claim.", "published_at": "2025-08-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PerPilot and the PerInstruct dataset for completing user-specific mobile instructions through memory retrieval, app exploration, and clarification when missing context remains unresolved.", "topics": ["memory", "planning", "human-in-the-loop", "mobile-navigation"], "updated_at": "2025-08-25", "url": "https://arxiv.org/abs/2508.18040"}
{"aliases": [], "authors": ["Haochen Pan", "Ryan Chard", "Reid Mello", "Christopher Grams", "Tanjin He", "Alexander Brace", "Owen Price Skelly", "Will Engler", "Hayden Holbrook", "Song Young Oh", "Maxime Gonthier", "Michael Papka", "Ben Blaiszik", "Kyle Chard", "Ian Foster"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "42485bfeab769c70dc852183958262ccbf0969619102b5d861525f868c865a9c", "submitted_at": "2025-08-25T21:02:33Z", "updated_at": "2025-08-25T21:02:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.18489v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.18489", "identifiers": {"arxiv": "2508.18489"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.18489"}], "name": "Experiences with Model Context Protocol Servers for Science and High Performance Computing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-25", "url": "https://arxiv.org/abs/2508.18489"}
{"aliases": [], "authors": ["Yuyang Zhao", "Wentao Shi", "Fuli Feng", "Xiangnan He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "02f4a76ae220caa289b322b622be6356086c5b04bc0f1598d0cd580c9ee25b6d", "submitted_at": "2025-08-26T05:23:24Z", "updated_at": "2025-08-27T04:25:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.18689v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.18689", "identifiers": {"arxiv": "2508.18689"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.18689"}], "name": "AppAgent-Pro: A Proactive GUI Agent System for Multidomain Information Integration and User Assistance", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes a proactive GUI assistant that gathers information across domains in response to user requests and anticipated information needs.", "topics": ["planning", "workflow"], "updated_at": "2025-08-27", "url": "https://arxiv.org/abs/2508.18689"}
{"aliases": [], "authors": ["Neil Kale", "Chen Bo Calvin Zhang", "Kevin Zhu", "Ankit Aich", "Paula Rodriguez", "Scale Red Team", "Christina Q. Knight", "Zifan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "72df2a296e56817a4bd0813d26f934254364a969a35c55f52496dbb793209ea6", "submitted_at": "2025-08-26T22:29:31Z", "updated_at": "2025-08-26T22:29:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.19461v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2508.19461", "identifiers": {"arxiv": "2508.19461"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.19461"}], "name": "Reliable Weak-to-Strong Monitoring of LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-26", "url": "https://arxiv.org/abs/2508.19461"}
{"aliases": [], "authors": ["Zhixin Lin", "Jungang Li", "Shidong Pan", "Yibo Shi", "Yue Yao", "Dongliang Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "815c8c1fa5993d2ef58798abd43a2b8945c476c27c777d30c1cbfacadd87af84", "submitted_at": "2025-08-27T00:41:28Z", "updated_at": "2025-09-03T07:48:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.19493v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2508.19493", "identifiers": {"arxiv": "2508.19493"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.19493"}], "name": "Mind the Third Eye! Benchmarking Privacy Awareness in MLLM-powered Smartphone Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SAPA-Bench with 7,138 smartphone scenarios to evaluate whether MLLM-powered mobile agents detect and protect sensitive personal information on screen.", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2025-09-03", "url": "https://arxiv.org/abs/2508.19493"}
{"aliases": [], "authors": ["Huaiyuan Yao", "Wanpeng Xu", "Justin Turnau", "Nadia Kellam", "Hua Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f4587857c480ea141e591504f117597a934054ec3f23bc943c743a2e0036d49", "submitted_at": "2025-08-27T06:45:06Z", "updated_at": "2026-01-31T00:09:08Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.19611v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}], "id": "arxiv:2508.19611", "identifiers": {"arxiv": "2508.19611"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.19611"}], "name": "Instructional Agents: Reducing Teaching Faculty Workload through Multi-Agent Instructional Design", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-31", "url": "https://arxiv.org/abs/2508.19611"}
{"aliases": [], "authors": ["Qihang Ai", "Pi Bu", "Yue Cao", "Yingyao Wang", "Jihao Gu", "Jingxuan Xing", "Zekun Zhu", "Wei Jiang", "Zhicheng Zheng", "Jun Song", "Yuning Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "385ed1b6e60c27660f0cbf281b0f2bbde052691801f7faae99cc053dfbc0d127", "submitted_at": "2025-08-27T08:40:05Z", "updated_at": "2026-04-25T13:29:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.19679v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2508.19679", "identifiers": {"arxiv": "2508.19679"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.19679"}], "name": "InquireMobile: Teaching VLM-based Mobile Agent to Request Human Assistance via Reinforcement Fine-Tuning", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-08-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "InquireMobile trains mobile agents to seek human confirmation at critical decisions using staged reinforcement fine-tuning and pre-action reasoning. InquireBench evaluates safe interaction and proactive inquiry, distinguishing asking for help from completing the underlying task.", "topics": ["human-in-the-loop", "reinforcement-learning", "evaluation", "mobile-navigation"], "updated_at": "2026-04-25", "url": "https://arxiv.org/abs/2508.19679"}
{"aliases": [], "authors": ["Zeyi Sun", "Yuhang Cao", "Jianze Liang", "Qiushi Sun", "Ziyu Liu", "Zhixiong Zhang", "Yuhang Zang", "Xiaoyi Dong", "Kai Chen", "Dahua Lin", "Jiaqi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac401f8d5c56050e72839f3bffba4f2c78bbdf34a00cab05fe504f03efd96a19", "submitted_at": "2025-08-27T17:59:50Z", "updated_at": "2025-08-27T17:59:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.20096v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.20096", "identifiers": {"arxiv": "2508.20096"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.20096"}], "name": "CODA: Coordinating the Cerebrum and Cerebellum for a Dual-Brain Computer Use Agent with Decoupled Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "CODA combines a generalist planner with a specialist GUI executor and trains application-specific planners through decoupled GRPO. Successful specialist trajectories then support supervised training of a consolidated planner, evaluated across four ScienceBoard applications.", "topics": ["reinforcement-learning", "planning", "science", "imitation-learning"], "updated_at": "2025-08-27", "url": "https://arxiv.org/abs/2508.20096"}
{"aliases": [], "authors": ["Sam Jung", "Agustin Garcinuno", "Spencer Mateega"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5b0143b84b6ebf66b8dbde0b1bea07d9eb22ca73542a0057a4a3af8b37170aa", "submitted_at": "2025-08-28T04:20:00Z", "updated_at": "2025-09-03T23:48:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.20410v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2508.20410", "identifiers": {"arxiv": "2508.20410"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.20410"}], "name": "UI-Bench: A Benchmark for Evaluating Design Capabilities of AI Text-to-App Tools", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-03", "url": "https://arxiv.org/abs/2508.20410"}
{"aliases": [], "authors": ["Mengxiao Wang", "Yuxuan Zhang", "Guofei Gu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64f50da9df01912b7f17e1c58be1974829cd3785c864b682e40f1a85ee51fd66", "submitted_at": "2025-08-28T15:19:07Z", "updated_at": "2025-09-16T01:18:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.20890v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2508.20890", "identifiers": {"arxiv": "2508.20890"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2508.20890"}], "name": "PromptSleuth: Detecting Prompt Injection via Semantic Intent Invariance", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-16", "url": "https://arxiv.org/abs/2508.20890"}
{"aliases": [], "authors": ["Yi-Hao Peng", "Dingzeyu Li", "Jeffrey P. Bigham", "Amy Pavel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5dc6411807fd58daed4824e86c39377b24d9fd022ac19daceb836bfb382353cf", "submitted_at": "2025-08-29T09:39:00Z", "updated_at": "2025-08-29T09:39:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.21456v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:40Z", "role": "primary", "sha256": "713ecfafe72365a57604306ccaba68f94158fc9b1aa169b408945922cdd1418c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2508.21456v1"}], "id": "arxiv:2508.21456", "identifiers": {"arxiv": "2508.21456"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:40Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.21456"}], "name": "Morae: Proactively Pausing UI Agents for User Choices", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-08-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Morae, a UI agent that detects decision points and pauses for user choices using interface code and screenshots. A study with blind and low-vision participants evaluates this mixed-initiative approach on real web tasks.", "topics": ["human-in-the-loop", "web-navigation", "evaluation"], "updated_at": "2025-08-29", "url": "https://arxiv.org/abs/2508.21456"}
{"aliases": [], "authors": ["Zhixiong Zeng", "Jing Huang", "Liming Zheng", "Wenkang Han", "Yufeng Zhong", "Lei Chen", "Longrong Yang", "Yingjie Chu", "Yuzhi He", "Lin Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b2a14819c7d7e8258947c9db8cc89012f4d133eaced4e4b904c3bfd1761181e5", "submitted_at": "2025-08-29T16:40:57Z", "updated_at": "2025-08-29T16:40:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.21767v1"}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2508.21767", "identifiers": {"arxiv": "2508.21767"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.21767"}], "name": "UItron: Foundational GUI Agent with Advanced Perception and Planning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "UItron combines GUI perception and planning data engineering with supervised training and curriculum reinforcement learning in connected mobile and PC environments. It also collects more than one million operation steps across 100 Chinese apps and builds offline and online evaluation environments.", "topics": ["data-generation", "reinforcement-learning", "planning", "grounding"], "updated_at": "2025-08-29", "url": "https://arxiv.org/abs/2508.21767"}
{"aliases": [], "authors": ["Ting-Chun Liu", "Ching-Yu Hsu", "Kuan-Yi Lee", "Chi-An Fu", "Hung-yi Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0edbeccaf18ae59dd8296274c5defe5bf0730f7c815668179737ff7611dfca80", "submitted_at": "2025-08-27T12:25:45Z", "updated_at": "2025-10-09T04:58:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.00088v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.00088", "identifiers": {"arxiv": "2509.00088"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.00088"}], "name": "AEGIS : Automated Co-Evolutionary Framework for Guarding Prompt Injections Schema", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-09", "url": "https://arxiv.org/abs/2509.00088"}
{"aliases": [], "authors": ["Ziyi Guan", "Jason Chun Lok Li", "Zhijian Hou", "Pingping Zhang", "Donglai Xu", "Yuzhi Zhao", "Mengyang Wu", "Jinpeng Chen", "Thanh-Toan Nguyen", "Pengfei Xian", "Wenao Ma", "Shengchao Qin", "Graziano Chesi", "Ngai Wong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c22f63bb5de22dd922ba37ea45df9404def017fee492e8ade03c83ab8b4f9b58", "submitted_at": "2025-08-30T05:32:32Z", "updated_at": "2025-08-30T05:32:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.00366v1"}}, "environments": ["android", "web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.00366", "identifiers": {"arxiv": "2509.00366"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.00366"}], "name": "KG-RAG: Enhancing GUI Agent Decision-Making via Knowledge Graph-Driven Retrieval-Augmented Generation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-08-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "KG-RAG converts UI transition graphs into a searchable vector store and uses intent-guided retrieval to construct actionable navigation paths. It introduces KG-Android-Bench and KG-Harmony-Bench and evaluates transfer from mobile apps to web and desktop tasks.", "topics": ["memory", "planning", "mobile-navigation", "evaluation"], "updated_at": "2025-08-30", "url": "https://arxiv.org/abs/2509.00366"}
{"aliases": [], "authors": ["Cheng Zhang", "Erhu Feng", "Xi Zhao", "Yisheng Zhao", "Wangbo Gong", "Jiahui Sun", "Dong Du", "Zhichao Hua", "Yubin Xia", "Haibo Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dbf26f28382245db085fd48f15e905bfa31f1c289a9a7f0799d35440069b442a", "submitted_at": "2025-08-30T15:24:47Z", "updated_at": "2025-08-30T15:24:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.00531v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:43Z", "role": "primary", "sha256": "b61937e1f046c4158b92c2f9d34dddea52a757f0e49b79f30efd73db24484abe", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2509.00531v1"}], "id": "arxiv:2509.00531", "identifiers": {"arxiv": "2509.00531"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.00531"}], "name": "MobiAgent: A Systematic Framework for Customizable Mobile Agents", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-08-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents MobiAgent, combining MobiMind agent models, the AgentRR acceleration framework, and MobiFlow evaluation. It also introduces an AI-assisted data-collection pipeline for mobile GUI tasks.", "topics": ["mobile-navigation", "efficiency", "data-generation"], "updated_at": "2025-08-30", "url": "https://arxiv.org/abs/2509.00531"}
{"aliases": [], "authors": ["Xiangchen Wang", "Jinrui Zhang", "Teng Wang", "Haigang Zhang", "Feng Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd30ee49237f62279edcb68be47768adad40a88f653674d5046069dd9096022b", "submitted_at": "2025-08-31T19:27:29Z", "updated_at": "2025-09-09T03:59:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.00969v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.00969", "identifiers": {"arxiv": "2509.00969"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.00969"}], "name": "Seeing More, Saying More: Lightweight Language Experts are Dynamic Video Token Compressors", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-09", "url": "https://arxiv.org/abs/2509.00969"}
{"aliases": [], "authors": ["Jaewoo Ahn", "Junseo Kim", "Heeseung Yun", "Jaehyeon Son", "Dongmin Park", "Jaewoong Cho", "Gunhee Kim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3bf86652aaeb252972f3010be38b75b78dfaa9a8063c480c80d0cea47397c2e7", "submitted_at": "2025-09-01T01:33:16Z", "updated_at": "2025-10-15T10:33:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.01052v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.01052", "identifiers": {"arxiv": "2509.01052"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.01052"}], "name": "FlashAdventure: A Benchmark for GUI Agents Solving Full Story Arcs in Diverse Adventure Games", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-01", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "FlashAdventure evaluates GUI agents on completing story arcs in 34 adventure games. It pairs an automated gameplay judge with COAST, a framework that keeps long-term clue memory to plan and complete sequential objectives.", "topics": ["evaluation", "long-horizon", "memory", "planning"], "updated_at": "2025-10-15", "url": "https://arxiv.org/abs/2509.01052"}
{"aliases": [], "authors": ["Abhinav Kumar", "Jaechul Roh", "Ali Naseh", "Amir Houmansadr", "Eugene Bagdasarian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f6ff68c59cede9cf59c039416a87ea39b926e3a3179b9f81af9354767befb93", "submitted_at": "2025-09-01T16:56:16Z", "updated_at": "2025-09-01T16:56:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.01619v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.01619", "identifiers": {"arxiv": "2509.01619"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.01619"}], "name": "Throttling Web Agents Using Reasoning Gates", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The contribution is an access-cost mechanism spanning websites and MCP servers, not a GUI navigation policy.", "published_at": "2025-09-01", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces reasoning gates that require agents to solve generated, verifiable puzzles before receiving access to resources. The throttling approach is deployed on a custom website and MCP servers and evaluated with web agents.", "topics": ["safety-security", "evaluation", "tool-integration"], "updated_at": "2025-09-01", "url": "https://arxiv.org/abs/2509.01619"}
{"aliases": [], "authors": ["Sherry Yang", "Joy He-Yueya", "Percy Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ade81f02014f5aa40ef642fd979530c97627f690d68cdf37d7c8fadf63ecdb98", "submitted_at": "2025-09-01T18:04:10Z", "updated_at": "2025-09-01T18:04:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.01684v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.01684", "identifiers": {"arxiv": "2509.01684"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.01684"}], "name": "Reinforcement Learning for Machine Learning Engineering Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-01", "url": "https://arxiv.org/abs/2509.01684"}
{"aliases": [], "authors": ["Georgios Papoudakis", "Thomas Coste", "Jianye Hao", "Jun Wang", "Kun Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "44c49412776ec9837a08d3d8347214495ffbda14a391c7fa67ad780a0eac9577", "submitted_at": "2025-09-01T18:55:27Z", "updated_at": "2025-11-13T01:26:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.01720v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2509.01720", "identifiers": {"arxiv": "2509.01720"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.01720"}], "name": "Succeed or Learn Slowly: Sample Efficient Off-Policy Reinforcement Learning for Mobile App Control", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Succeed or Learn Slowly applies direct off-policy updates to successful interactions and more conservative regularized updates to negative examples. Successful Transition Replay prioritizes successful experience during policy learning, with evaluation on AndroidWorld.", "topics": ["reinforcement-learning", "mobile-navigation", "efficiency"], "updated_at": "2025-11-13", "url": "https://arxiv.org/abs/2509.01720"}
{"aliases": [], "authors": ["Longrong Yang", "Zhixiong Zeng", "Yufeng Zhong", "Jing Huang", "Liming Zheng", "Lei Chen", "Haibo Qiu", "Zequn Qin", "Lin Ma", "Xi Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "227630b3a257557262ed25eb21906a129eb35c6c5d2de7c5754f919dd281c4bd", "submitted_at": "2025-09-02T13:47:54Z", "updated_at": "2025-09-02T13:47:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.02322v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2509.02322", "identifiers": {"arxiv": "2509.02322"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.02322"}], "name": "OmniActor: A Generalist GUI and Embodied Agent for 2D&3D Worlds", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-09-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces OmniActor, which shares shallow model layers while separating deeper parameters to train jointly on GUI and embodied interactions. A unified action representation and mixed training data support evaluation in both task families.", "topics": ["action-prediction", "data-generation", "imitation-learning"], "updated_at": "2025-09-02", "url": "https://arxiv.org/abs/2509.02322"}
{"aliases": [], "authors": ["Jingru Fan", "Yufan Dang", "Jingyao Wu", "Huatao Li", "Runde Yang", "Xiyuan Yang", "Yuheng Wang", "Chen Qian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49a887b12b3395821aa100a66e61aac808309ef947ae752475df3d2db9a4fcd8", "submitted_at": "2025-09-02T15:48:21Z", "updated_at": "2025-10-17T00:57:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.02444v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:33.619660+00:00", "role": "primary", "sha256": "6c2591dad8893f87cffdea211584a3733dc2137dafa9416c1885d1b0126fd540", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2509.02444v2"}], "id": "arxiv:2509.02444", "identifiers": {"arxiv": "2509.02444"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:10.946679+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.02444"}], "name": "AppCopilot: Toward General, Accurate, Long-Horizon, and Efficient Mobile Agent", "notes": "Trace reuse assumes stable layouts and fixed tasks. The described function-call routes are user-selected, and coordinate calibration depends on valid detected widgets. No particular operating-system coverage, latency guarantee or reproduced result is asserted. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (4.2 OCR/OR Region Localization; 4.2.2 Coordinate Calibration; 4.3 Multi-Agent Voting; 5.2 Task Decomposition; 6.2 Experience Reuse; 6.3 Function Calling); not an entire-paper reading or reproduction claim.", "published_at": "2025-09-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a mobile-agent stack combining region-based coordinate correction, hierarchical task planning, multi-agent voting, and reuse of successful interaction traces.", "topics": ["planning", "grounding", "workflow", "mobile-navigation", "long-horizon"], "updated_at": "2025-10-17", "url": "https://arxiv.org/abs/2509.02444"}
{"aliases": [], "authors": ["Ziting Wang", "Shize Zhang", "Haitao Yuan", "Jinwei Zhu", "Wei Dong", "Gao Cong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1850899f0cb84621c69fe11ab13f2ba0997afb8ad5541ca47d29edaa2400839a", "submitted_at": "2025-09-02T16:25:12Z", "updated_at": "2026-08-15T04:45:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.02473v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2509.02473", "identifiers": {"arxiv": "2509.02473"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.02473"}], "name": "FDABench: A Benchmark for Data Agents on Analytical Queries over Heterogeneous Data", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-15", "url": "https://arxiv.org/abs/2509.02473"}
{"aliases": [], "authors": ["Haoming Wang", "Haoyang Zou", "Huatong Song", "Jiazhan Feng", "Junjie Fang", "Junting Lu", "Longxiang Liu", "Qinyu Luo", "Shihao Liang", "Shijue Huang", "Wanjun Zhong", "Yining Ye", "Yujia Qin", "Yuwen Xiong", "Yuxin Song", "Zhiyong Wu", "Aoyan Li", "Bo Li", "Chen Dun", "Chong Liu", "Daoguang Zan", "Fuxing Leng", "Hanbin Wang", "Hao Yu", "Haobin Chen", "Hongyi Guo", "Jing Su", "Jingjia Huang", "Kai Shen", "Kaiyu Shi", "Lin Yan", "Peiyao Zhao", "Pengfei Liu", "Qinghao Ye", "Renjie Zheng", "Shulin Xin", "Wayne Xin Zhao", "Wen Heng", "Wenhao Huang", "Wenqian Wang", "Xiaobo Qin", "Yi Lin", "Youbin Wu", "Zehui Chen", "Zihao Wang", "Baoquan Zhong", "Xinchun Zhang", "Xujing Li", "Yuanfan Li", "Zhongkai Zhao", "Chengquan Jiang", "Faming Wu", "Haotian Zhou", "Jinlin Pang", "Li Han", "Qi Liu", "Qianli Ma", "Siyao Liu", "Songhua Cai", "Wenqi Fu", "Xin Liu", "Yaohui Wang", "Zhi Zhang", "Bo Zhou", "Guoliang Li", "Jiajun Shi", "Jiale Yang", "Jie Tang", "Li Li", "Qihua Han", "Taoran Lu", "Woyu Lin", "Xiaokang Tong", "Xinyao Li", "Yichi Zhang", "Yu Miao", "Zhengxuan Jiang", "Zili Li", "Ziyuan Zhao", "Chenxin Li", "Dehua Ma", "Feng Lin", "Ge Zhang", "Haihua Yang", "Hangyu Guo", "Hongda Zhu", "Jiaheng Liu", "Junda Du", "Kai Cai", "Kuanye Li", "Lichen Yuan", "Meilan Han", "Minchao Wang", "Shuyue Guo", "Tianhao Cheng", "Xiaobo Ma", "Xiaojun Xiao", "Xiaolong Huang", "Xinjie Chen", "Yidi Du", "Yilin Chen", "Yiwen Wang", "Zhaojian Li", "Zhenzhu Yang", "Zhiyuan Zeng", "Chaolin Jin", "Chen Li", "Hao Chen", "Haoli Chen", "Jian Chen", "Qinghao Zhao", "Guang Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "800e417f4a2f720beb436ddd2e7bc60f0dd2483b44ac007a72a5eab2c6001ff8", "submitted_at": "2025-09-02T17:44:45Z", "updated_at": "2025-09-05T14:59:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.02544v2"}}, "environments": ["web", "desktop", "windows", "android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.02544", "identifiers": {"arxiv": "2509.02544"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.02544"}], "name": "UI-TARS-2 Technical Report: Advancing GUI Agent with Multi-Turn Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-TARS-2 combines scalable trajectory generation, stabilized multi-turn reinforcement learning, and a sandbox that supports GUI, filesystem, and terminal actions. The report evaluates the resulting agent across web, desktop, Android, and game tasks and analyzes training stability.", "topics": ["reinforcement-learning", "data-generation", "sandbox", "tool-integration"], "updated_at": "2025-09-05", "url": "https://arxiv.org/abs/2509.02544"}
{"aliases": [], "authors": ["Weizhi Chen", "Ziwei Wang", "Leyang Yang", "Sheng Zhou", "Xiaoxuan Tang", "Jiajun Bu", "Yong Li", "Wei Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97bc25c43e1e431cb9db4cdc0ca1e94c65f0906438a8a627d5d2bbb2d92b8460", "submitted_at": "2025-08-27T12:31:37Z", "updated_at": "2025-08-27T12:31:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.03536v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.03536", "identifiers": {"arxiv": "2509.03536"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.03536"}], "name": "PG-Agent: An Agent Powered by Page Graph", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-08-27", "url": "https://arxiv.org/abs/2509.03536"}
{"aliases": [], "authors": ["Gowen Loo", "Chang Liu", "Qinghong Yin", "Xiang Chen", "Jiawei Chen", "Jingyuan Zhang", "Yu Tian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8b2db00f62afe350e55e2266b3627de9b0cf089178b0cf80eb13b06275eee15", "submitted_at": "2025-09-04T05:22:42Z", "updated_at": "2025-09-04T05:22:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.03891v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2509.03891", "identifiers": {"arxiv": "2509.03891"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.03891"}], "name": "MobileRAG: Enhancing Mobile Agent with Retrieval-Augmented Generation", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileRAG combines retrieval and memory components to support mobile tasks that need external knowledge or long action sequences. MobileRAG-Eval is introduced to study these requirements on complex real-world app tasks.", "topics": ["memory", "mobile-navigation", "evaluation", "long-horizon"], "updated_at": "2025-09-04", "url": "https://arxiv.org/abs/2509.03891"}
{"aliases": [], "authors": ["Yifan Yang", "Zhixiang Duan", "Tianshi Xie", "Fuyu Cao", "Pinxi Shen", "Peili Song", "Piaopiao Jin", "Guokang Sun", "Shaoqing Xu", "Yangwei You", "Jingtai Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a71b0fbc2947157bd4d4373dce5ca8c4a9c6560de94680f54a0dae6fbb1aaa9", "submitted_at": "2025-09-04T08:47:26Z", "updated_at": "2025-12-03T06:13:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.04018v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.04018", "identifiers": {"arxiv": "2509.04018"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.04018"}], "name": "FPC-VLA: A Vision-Language-Action Framework with a Supervisor for Failure Prediction and Correction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-03", "url": "https://arxiv.org/abs/2509.04018"}
{"aliases": [], "authors": ["Petr Průcha", "Michaela Matoušková", "Jan Strnad"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9d2c3086aac59964b4c34a0ba890e6a9a1e55cf45c1af604f1f856180a5db746", "submitted_at": "2025-09-04T13:22:44Z", "updated_at": "2025-09-04T13:22:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.04198v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2509.04198", "identifiers": {"arxiv": "2509.04198"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.04198"}], "name": "Are LLM Agents the New RPA? A Comparative Study with RPA Across Enterprise Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-04", "url": "https://arxiv.org/abs/2509.04198"}
{"aliases": [], "authors": ["Wanfu Wang", "Qipeng Huang", "Guangquan Xue", "Xiaobo Liang", "Juntao Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "875e2452985bae9e67d7071121ab13f5205307ecc30bf33f237048a44e75b3bc", "submitted_at": "2025-09-04T14:17:01Z", "updated_at": "2025-09-04T14:17:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.04243v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.04243", "identifiers": {"arxiv": "2509.04243"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.04243"}], "name": "Learning Active Perception via Self-Evolving Preference Optimization for GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "LASER builds preference data for multi-step GUI perception by combining Monte Carlo quality estimates with region-overlap evaluation. Its self-evolving training encourages relevant region selection and adapts the amount of reasoning before predicting screen coordinates.", "topics": ["grounding", "data-generation", "screen-understanding"], "updated_at": "2025-09-04", "url": "https://arxiv.org/abs/2509.04243"}
{"aliases": [], "authors": ["Hongyi Jing", "Jiafu Chen", "Chen Rao", "Ziqiang Dang", "Jiajie Teng", "Tianyi Chu", "Juncheng Mo", "Shuo Fang", "Huaizhong Lin", "Rui Lv", "Chenguang Ma", "Lei Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f74081d31c98d18e10a5febc5de436a3788fc4156215e4be8fc54b1c53260f58", "submitted_at": "2025-09-05T08:24:12Z", "updated_at": "2025-09-05T08:24:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.04908v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:46Z", "role": "primary", "sha256": "e4304967a9a63ead71f00e1933965321540df913f534c0711cc18b9190560cde", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2509.04908v1"}], "id": "arxiv:2509.04908", "identifiers": {"arxiv": "2509.04908"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:46Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.04908"}], "name": "SparkUI-Parser: Enhancing GUI Perception with Robust Grounding and Parsing", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-09-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SparkUI-Parser, which adds continuous coordinate decoding and a rejection mechanism for nonexistent targets to a multimodal model. It also presents ScreenParse for evaluating structural interface perception.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2025-09-05", "url": "https://arxiv.org/abs/2509.04908"}
{"aliases": [], "authors": ["Quan Chen", "Chenrui Shi", "Qi Chen", "Yuwei Wu", "Zhi Gao", "Xintong Zhang", "Rui Gao", "Kun Wu", "Yunde Jia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f09361ea5957cc7f1589962aedae9f34662b7d7a96833b194928f14cc5d5897a", "submitted_at": "2025-09-04T06:10:32Z", "updated_at": "2025-12-18T08:12:36Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.05368v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.05368", "identifiers": {"arxiv": "2509.05368"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.05368"}], "name": "Long-Horizon Visual Imitation Learning via Plan and Code Reflection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-18", "url": "https://arxiv.org/abs/2509.05368"}
{"aliases": [], "authors": ["Ishaan Verma", "Arsheya Yadav"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "56d8bd92ce390e28bab71411440f981ab9a14a77dc4530ec19b67eac0ee4e102", "submitted_at": "2025-09-06T21:05:18Z", "updated_at": "2025-11-11T06:53:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.05831v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.05831", "identifiers": {"arxiv": "2509.05831"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.05831"}], "name": "Decoding Latent Attack Surfaces in LLMs: Prompt Injection via HTML in Web Summarization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-11", "url": "https://arxiv.org/abs/2509.05831"}
{"aliases": [], "authors": ["Pengxiang Zhao", "Guangyi Liu", "YaoZhen Liang", "Weiqing He", "Zhengxi Lu", "WenHao Wang", "Yuehao Huang", "Yuxiang Chai", "Zhaolu Kang", "Yaxuan Guo", "Hao Wang", "Kexin Zhang", "Liang Liu", "Yong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b1e0e807f5f396bfd9108ab6cfacd48ed4780abfb792589da0b41d7d246c717", "submitted_at": "2025-09-08T09:43:48Z", "updated_at": "2026-04-15T11:26:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.06477v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.06477", "identifiers": {"arxiv": "2509.06477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.06477"}], "name": "MAS-Bench: A Unified Benchmark for Shortcut-Augmented Hybrid Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "MAS-Bench evaluates mobile agents that combine GUI actions with APIs, deep links, and reusable scripts. Its 139 tasks across 11 apps test use of 88 predefined shortcuts as well as the ability to generate new reusable workflows.", "topics": ["evaluation", "mobile-navigation", "tool-integration", "workflow"], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2509.06477"}
{"aliases": [], "authors": ["Junteng Liu", "Yunji Li", "Chi Zhang", "Jingyang Li", "Aili Chen", "Ke Ji", "Weiyu Cheng", "Zijia Wu", "Chengyu Du", "Qidi Xu", "Jiayuan Song", "Zhengmao Zhu", "Wenhu Chen", "Pengyu Zhao", "Junxian He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "120d73f67e9e530b252df457e6157bba1d0a520a6825b9101db41fe65f9b83b7", "submitted_at": "2025-09-08T10:07:03Z", "updated_at": "2025-09-26T02:31:59Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.06501v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.06501", "identifiers": {"arxiv": "2509.06501"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.06501"}], "name": "WebExplorer: Explore and Evolve for Training Long-Horizon Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-26", "url": "https://arxiv.org/abs/2509.06501"}
{"aliases": [], "authors": ["Jiacheng Miao", "Joe R. Davis", "Yaohui Zhang", "Jonathan K. Pritchard", "James Zou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4843287de3408f865e3c44a1ab5c0447b95f72447000cd77eab041d768d780ae", "submitted_at": "2025-09-08T17:28:42Z", "updated_at": "2025-10-16T14:09:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.06917v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.06917", "identifiers": {"arxiv": "2509.06917"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.06917"}], "name": "Paper2Agent: Reimagining Research Papers As Interactive and Reliable AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-16", "url": "https://arxiv.org/abs/2509.06917"}
{"aliases": [], "authors": ["Yinheng Li", "Hailey Hultquist", "Justin Wagle", "Kazuhito Koishida"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f5e5d68ea5cfe04d3e4bdb31d3b6152d44db3de344ea7afa950b5400a9dba5c", "submitted_at": "2025-09-08T18:00:12Z", "updated_at": "2025-09-08T18:00:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.07098v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.07098", "identifiers": {"arxiv": "2509.07098"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.07098"}], "name": "Instruction Agent: Enhancing Agent with Expert Demonstration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-08", "url": "https://arxiv.org/abs/2509.07098"}
{"aliases": [], "authors": ["Zheng Wu", "Heyuan Huang", "Xingyu Lou", "Xiangmou Qu", "Pengzhou Cheng", "Zongru Wu", "Weiwen Liu", "Weinan Zhang", "Jun Wang", "Zhaoxiang Wang", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3bd6b323a376d7606b8c5ec0615b992f47a388cbd3604191241f6a0a9b2cce6e", "submitted_at": "2025-09-09T09:46:01Z", "updated_at": "2026-04-03T02:36:28Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.07553v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:50Z", "role": "primary", "sha256": "776192081826494b05077d71757f3006575ed50a8cc244004e2de7053104d7eb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2509.07553v3"}], "id": "arxiv:2509.07553", "identifiers": {"arxiv": "2509.07553"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.07553"}], "name": "VeriOS: Query-Driven Proactive Human-Agent-GUI Interaction for Trustworthy OS Agents", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-09-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces VeriOS-Agent, which learns when to execute GUI actions autonomously and when to query a user under untrustworthy conditions. Its staged training combines supervised learning and group-relative policy optimization.", "topics": ["human-in-the-loop", "safety-security", "reinforcement-learning"], "updated_at": "2026-04-03", "url": "https://arxiv.org/abs/2509.07553"}
{"aliases": [], "authors": ["Haitao Hu", "Peng Chen", "Yanpeng Zhao", "Yuqi Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae582aceb31075c460a9e4a6dce9978574c3e0ada731b26c13738a4a4fafcc32", "submitted_at": "2025-09-09T13:59:00Z", "updated_at": "2025-09-09T13:59:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.07764v1"}}, "environments": ["desktop", "linux", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "topics", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:25:42Z", "role": "primary", "sha256": "7f303629ef6d9c7df128c6d12d472abd9bf238e0204515f18839249adef317b9", "supports": ["summary", "environments", "scope", "notes"], "url": "https://arxiv.org/html/2509.07764v1"}], "id": "arxiv:2509.07764", "identifiers": {"arxiv": "2509.07764"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.07764"}], "name": "AgentSentinel: An End-to-End and Real-Time Security Defense Framework for Computer-Use Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The official full paper confirms GUI/desktop interaction alongside Bash and file tools. Its abstract and body enumerate attack categories differently; no category count is promoted into this annotation.", "published_at": "2025-09-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AgentSentinel, which intercepts sensitive computer-agent operations and audits them using task context together with system traces. BadComputerUse evaluates this enforcement approach across attack scenarios involving tool and computer control.", "topics": ["safety-security", "observability", "evaluation"], "updated_at": "2025-09-09", "url": "https://arxiv.org/abs/2509.07764"}
{"aliases": [], "authors": ["Zongzheng Zhang", "Chenghao Yue", "Haobo Xu", "Minwen Liao", "Xianglin Qi", "Huan-ang Gao", "Ziwei Wang", "Hao Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97eaa15391dc08fbc97063e6e76b7bae725a05b3762d61963757d5d24fa0552d", "submitted_at": "2025-09-10T17:52:09Z", "updated_at": "2025-09-10T17:52:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.08820v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.08820", "identifiers": {"arxiv": "2509.08820"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.08820"}], "name": "RoboChemist: Long-Horizon and Safety-Compliant Robotic Chemical Experimentation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-10", "url": "https://arxiv.org/abs/2509.08820"}
{"aliases": [], "authors": ["Hangyi Jia", "Yuxi Qian", "Hanwen Tong", "Xinhui Wu", "Lin Chen", "Feng Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f17328cfdaacd90ec9578cb4168b1614c60c176d47f7b74012f188f7438882b0", "submitted_at": "2025-09-11T10:10:48Z", "updated_at": "2025-09-11T10:10:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.09321v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.09321", "identifiers": {"arxiv": "2509.09321"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.09321"}], "name": "Towards Adaptive ML Benchmarks: Web-Agent-Driven Construction, Domain Expansion, and Metric Optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-11", "url": "https://arxiv.org/abs/2509.09321"}
{"aliases": [], "authors": ["Georgios Pantazopoulos", "Eda B. Özyiğit"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6ea81a5e70bc9ed0aa9143c847035dd6bf1abf04314e322e28d9f732e655dd23", "submitted_at": "2025-09-12T15:33:49Z", "updated_at": "2025-09-15T08:46:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.10345v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.10345", "identifiers": {"arxiv": "2509.10345"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.10345"}], "name": "Towards Understanding Visual Grounding in Visual Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-15", "url": "https://arxiv.org/abs/2509.10345"}
{"aliases": [], "authors": ["Pavan Reddy", "Aditya Sanjay Gujral"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1d615686ac83b81114ddcf69d984ffc955279413d2062c66c500ca10566dff4b", "submitted_at": "2025-09-06T04:06:01Z", "updated_at": "2025-09-06T04:06:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.10540v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.10540", "identifiers": {"arxiv": "2509.10540"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.10540"}], "name": "EchoLeak: The First Real-World Zero-Click Prompt Injection Exploit in a Production LLM System", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-06", "url": "https://arxiv.org/abs/2509.10540"}
{"aliases": [], "authors": ["Jingyu Tang", "Chaoran Chen", "Jiawen Li", "Zhiping Zhang", "Bingcan Guo", "Ibrahim Khalilov", "Simret Araya Gebreegziabher", "Bingsheng Yao", "Dakuo Wang", "Yanfang Ye", "Tianshi Li", "Ziang Xiao", "Yaxing Yao", "Toby Jia-Jun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "247ed053a4c8f62106d9f8e9af0801d94265ed712f4ae9f095fd8aaccbe33c59", "submitted_at": "2025-09-12T22:26:31Z", "updated_at": "2025-09-12T22:26:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.10723v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.10723", "identifiers": {"arxiv": "2509.10723"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.10723"}], "name": "Dark Patterns Meet GUI Agents: LLM Agent Susceptibility to Manipulative Interfaces and the Role of Human Oversight", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-09-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies how GUI agents, people, and human-agent teams respond to manipulative interface designs. The two-phase evaluation examines failure modes and the benefits and costs of human oversight.", "topics": ["human-in-the-loop", "safety-security", "evaluation"], "updated_at": "2025-09-12", "url": "https://arxiv.org/abs/2509.10723"}
{"aliases": [], "authors": ["Yitong Zhang", "Ximo Li", "Liyi Cai", "Jia Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ab96edd8e52b3a490455b46878c92c23d15c0f9e105e71dbd6848e058da42ec", "submitted_at": "2025-09-14T12:47:54Z", "updated_at": "2026-08-24T12:04:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.11250v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.11250", "identifiers": {"arxiv": "2509.11250"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.11250"}], "name": "Environmental Injection Attacks against GUI Agents in Realistic Dynamic Environments", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-09-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Chameleon, an environmental-injection evaluation framework that varies webpage content and trigger context. It combines simulated page variation with attention-based optimization and tests GUI agents across realistic websites.", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2026-08-24", "url": "https://arxiv.org/abs/2509.11250"}
{"aliases": [], "authors": ["Xiao Liang", "Bangxin Li", "Zixuan Chen", "Hanyue Zheng", "Zhi Ma", "Di Wang", "Cong Tian", "Quan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0b2fca1d3d96e2e85d875333b45c0ce99c80e5e5ae33faf90a242f7eefe26817", "submitted_at": "2025-09-14T12:54:21Z", "updated_at": "2026-04-21T04:59:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.11253v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.11253", "identifiers": {"arxiv": "2509.11253"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.11253"}], "name": "VideoAgent: Personalized Synthesis of Scientific Videos", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-21", "url": "https://arxiv.org/abs/2509.11253"}
{"aliases": [], "authors": ["Zhengxi Lu", "Jiabo Ye", "Fei Tang", "Yongliang Shen", "Haiyang Xu", "Ziwei Zheng", "Weiming Lu", "Ming Yan", "Fei Huang", "Jun Xiao", "Yueting Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1e31638e531aff49fe59bef6e80b35069b1f851ea0cbf1901130557148df4783", "submitted_at": "2025-09-15T03:24:08Z", "updated_at": "2025-09-24T15:05:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.11543v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "id": "arxiv:2509.11543", "identifiers": {"arxiv": "2509.11543"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.11543"}], "name": "UI-S1: Advancing GUI Automation via Semi-online Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-S1 simulates multi-turn online learning over offline trajectories, using a patch module to recover divergences while retaining the model’s own outputs in context. It combines discounted future returns with step- and episode-level advantages and introduces a semi-online performance proxy.", "topics": ["reinforcement-learning", "evaluation", "mobile-navigation"], "updated_at": "2025-09-24", "url": "https://arxiv.org/abs/2509.11543"}
{"aliases": [], "authors": ["Weiming Li", "Yan Shao", "Jing Yang", "Yujing Lu", "Ling Zhong", "Yuhan Wang", "Min Yu", "Tongxiao Ruan", "Manni Duan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75ffc21abebbc2c362798f3455b9e6240b58407572e5633d5171e5612cfae71b", "submitted_at": "2025-09-15T03:28:29Z", "updated_at": "2026-06-10T01:13:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.11548v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.11548", "identifiers": {"arxiv": "2509.11548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.11548"}], "name": "How Auxiliary Reasoning Unleashes GUI Grounding in VLMs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies the gap between latent visual grounding and explicit coordinate output in general vision-language models. Three zero-shot methods add axes, grids, or labeled intersections to GUI screenshots so models can express spatial locations without fine-tuning.", "topics": ["grounding", "training-free", "screen-understanding"], "updated_at": "2026-06-10", "url": "https://arxiv.org/abs/2509.11548"}
{"aliases": [], "authors": ["Shuning Zhang", "Yutong Jiang", "Rongjun Ma", "Yuting Yang", "Mingyao Xu", "Zhixin Huang", "Xin Yi", "Hewu Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f37ebc10535f96d17d28990e608a1f2eb6cc004f40d58b3288c4a81d781160ea", "submitted_at": "2025-09-15T13:58:52Z", "updated_at": "2025-09-15T13:58:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.11939v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.11939", "identifiers": {"arxiv": "2509.11939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.11939"}], "name": "PrivWeb: Unobtrusive and Content-aware Privacy Protection For Web Agents", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-09-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PrivWeb, which uses a local language model to anonymize interface information according to user preferences. Adaptive notifications and selective pauses give users control over sensitive information access during web-agent tasks.", "topics": ["safety-security", "human-in-the-loop", "web-navigation"], "updated_at": "2025-09-15", "url": "https://arxiv.org/abs/2509.11939"}
{"aliases": [], "authors": ["Harshit Rajgarhia", "Shivali Dalmia", "Mengyang Zhao", "Mukherji Abhishek", "Kiran Ganesh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89cdd479be62d998e0cbc01946e4b69ef98a85ffac59037522534b7f40ddd839", "submitted_at": "2025-09-16T00:52:41Z", "updated_at": "2025-10-06T21:30:41Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.12543v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.12543", "identifiers": {"arxiv": "2509.12543"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.12543"}], "name": "Human + AI for Accelerating Ad Localization Evaluation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-06", "url": "https://arxiv.org/abs/2509.12543"}
{"aliases": [], "authors": ["Yarin Benyamin", "Argaman Mordoch", "Shahaf S. Shperberg", "Roni Stern"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "647578f622e6166901349d0477095956bb7b8b85e1336ef5e6871872010a27f9", "submitted_at": "2025-09-16T11:51:07Z", "updated_at": "2025-09-16T11:51:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.12987v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.12987", "identifiers": {"arxiv": "2509.12987"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.12987"}], "name": "Toward PDDL Planning Copilot", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-16", "url": "https://arxiv.org/abs/2509.12987"}
{"aliases": [], "authors": ["Zongru Wu", "Rui Mao", "Zhiyuan Tian", "Pengzhou Cheng", "Tianjie Ju", "Zheng Wu", "Lingzhong Dong", "Haiyue Sheng", "Zhuosheng Zhang", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "be8620b021bd9218c1143282f1c7a7ea06be2abe6fc0928a310cfc8ac3172a18", "submitted_at": "2025-09-17T01:14:14Z", "updated_at": "2026-03-18T08:12:54Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2509.13615v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:53Z", "role": "primary", "sha256": "28426488ebca15b74a693ea9fcb2b0ce3165d3baf97b76a4884b34a0d1997af4", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2509.13615v5"}], "id": "arxiv:2509.13615", "identifiers": {"arxiv": "2509.13615"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:53Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.13615"}], "name": "See, Think, Act: Teaching Multimodal Agents to Effectively Interact with GUI by Identifying Toggles", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-09-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a benchmark for binary GUI-toggle instructions and State-aware Reasoning, which compares the observed control state with the requested state before choosing an action. It evaluates agents on toggle tasks and broader interactive benchmarks.", "topics": ["state-tracking", "grounding", "evaluation"], "updated_at": "2026-03-18", "url": "https://arxiv.org/abs/2509.13615"}
{"aliases": [], "authors": ["Liangtao Lin", "Zhaomeng Zhu", "Tianwei Zhang", "Yonggang Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c1a4019df77278943829e16b155771abb1c8ca30726044cf5969227e0437f5f", "submitted_at": "2025-09-17T05:14:11Z", "updated_at": "2025-09-17T05:14:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.13704v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.13704", "identifiers": {"arxiv": "2509.13704"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.13704"}], "name": "InfraMind: A Novel Exploration-based GUI Agentic Framework for Mission-critical Industrial Management", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "InfraMind explores industrial management GUIs using virtual-machine snapshots and builds memory for planning and hierarchical state recognition. Its framework also studies knowledge distillation and safety checks on open-source and commercial data-center management platforms.", "topics": ["industry", "planning", "memory", "safety-security"], "updated_at": "2025-09-17", "url": "https://arxiv.org/abs/2509.13704"}
{"aliases": [], "authors": ["Ziyuan Chen", "Zhenghui Zhao", "Zhangye Han", "Miancan Liu", "Xianhang Ye", "Yiqing Li", "Hongbo Min", "Jinkui Ren", "Xiantao Zhang", "Guitao Cao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc1e4f78f1d3ffdb1b67f7c0bb645561a76db19dfa5f442fa9aa0a0d4d1ae0de", "submitted_at": "2025-09-17T16:58:44Z", "updated_at": "2025-09-19T02:13:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.14172v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.14172", "identifiers": {"arxiv": "2509.14172"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.14172"}], "name": "TGPO: Tree-Guided Preference Optimization for Robust Web Agent Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "TGPO merges semantically equivalent web states into tree-structured trajectories for offline preference optimization. A process reward model scores subgoal progress, redundant behavior, and action validity, while dynamic weighting emphasizes consequential decisions.", "topics": ["reinforcement-learning", "reward-models", "web-navigation"], "updated_at": "2025-09-19", "url": "https://arxiv.org/abs/2509.14172"}
{"aliases": [], "authors": ["Daniel Röder", "Akhil Juneja", "Roland Roller", "Sven Schmeier"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac93e8cfca06ff86fad06408051abaa6b6510a44d7ad9f2d34c478b11391a32f", "submitted_at": "2025-09-17T19:34:49Z", "updated_at": "2025-09-17T19:34:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.14382v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.14382", "identifiers": {"arxiv": "2509.14382"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.14382"}], "name": "Detecting Pipeline Failures through Fine-Grained Analysis of Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a modular framework for identifying errors at intermediate stages of a web-agent pipeline. A SeeAct and Mind2Web case study examines weaknesses that are obscured by final task-success metrics.", "topics": ["evaluation", "observability", "web-navigation"], "updated_at": "2025-09-17", "url": "https://arxiv.org/abs/2509.14382"}
{"aliases": [], "authors": ["Yi Lin", "Lujin Zhao", "Yijie Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "20b7fcd1f51d3625533474a7fcea37118a4833019824135a75e352ee030c36a1", "submitted_at": "2025-09-18T02:24:14Z", "updated_at": "2025-09-18T02:24:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.14547v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.14547", "identifiers": {"arxiv": "2509.14547"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.14547"}], "name": "(P)rior(D)yna(F)low: A Priori Dynamic Workflow Construction via Multi-Agent Collaboration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-18", "url": "https://arxiv.org/abs/2509.14547"}
{"aliases": [], "authors": ["Worawalan Chatlatanagulchai", "Kundjanasith Thonglek", "Brittany Reid", "Yutaro Kashiwa", "Pattara Leelaprute", "Arnon Rungsawang", "Bundit Manaskasemsak", "Hajimu Iida"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b45af31e36cf68ca7147fb3da4a1aaf0151a44241ed18befd9c77939a8b31102", "submitted_at": "2025-09-18T08:46:41Z", "updated_at": "2025-09-18T08:46:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.14744v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.14744", "identifiers": {"arxiv": "2509.14744"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.14744"}], "name": "On the Use of Agentic Coding Manifests: An Empirical Study of Claude Code", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-18", "url": "https://arxiv.org/abs/2509.14744"}
{"aliases": [], "authors": ["Xingyao Lin", "Xinghao Zhu", "Tianyi Lu", "Guojin Zhong", "Sicheng Xie", "Hui Zhang", "Xipeng Qiu", "Zuxuan Wu", "Yu-Gang Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb954c487436e1031507419a4d4d23b5631108651594dbe6b791fc4389c90eeb", "submitted_at": "2025-09-18T15:25:31Z", "updated_at": "2026-07-03T09:06:35Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.15061v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.15061", "identifiers": {"arxiv": "2509.15061"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.15061"}], "name": "Ask-to-Clarify: Resolving Instruction Ambiguity through Multi-turn Dialogue", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-03", "url": "https://arxiv.org/abs/2509.15061"}
{"aliases": [], "authors": ["Zhaoyang Liu", "Jingjing Xie", "Zichen Ding", "Zehao Li", "Bowen Yang", "Zhenyu Wu", "Xuehui Wang", "Qiushi Sun", "Shi Liu", "Weiyun Wang", "Shenglong Ye", "Qingyun Li", "Xuan Dong", "Yue Yu", "Chenyu Lu", "YunXiang Mo", "Yao Yan", "Zeyue Tian", "Xiao Zhang", "Yuan Huang", "Yiqian Liu", "Weijie Su", "Gen Luo", "Xiangyu Yue", "Biqing Qi", "Kai Chen", "Bowen Zhou", "Yu Qiao", "Qifeng Chen", "Wenhai Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "70d09da7d4990118b2b0745cbddf2bab25852ba3ec26625902fee94cdab37c8d", "submitted_at": "2025-09-18T17:59:22Z", "updated_at": "2025-09-19T05:29:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.15221v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.15221", "identifiers": {"arxiv": "2509.15221"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.15221"}], "name": "ScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a cross-platform computer-use dataset and agents trained with a collection pipeline combining automated interaction and human input.", "topics": ["data-generation", "automation"], "updated_at": "2025-09-19", "url": "https://arxiv.org/abs/2509.15221"}
{"aliases": [], "authors": ["Di Wen", "Kunyu Peng", "Junwei Zheng", "Yufan Chen", "Yitian Shi", "Jiale Wei", "Ruiping Liu", "Kailun Yang", "Rainer Stiefelhagen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8cd5bcf8e19e7fc8e413c7cad7ff7f9751897625a022848aad6d9da90c2cfd03", "submitted_at": "2025-09-17T14:36:38Z", "updated_at": "2026-03-09T02:29:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.15237v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.15237", "identifiers": {"arxiv": "2509.15237"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.15237"}], "name": "MICA: Multi-Agent Industrial Coordination Assistant", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-09", "url": "https://arxiv.org/abs/2509.15237"}
{"aliases": [], "authors": ["Xianhang Ye", "Yiqing Li", "Wei Dai", "Miancan Liu", "Ziyuan Chen", "Zhangye Han", "Hongbo Min", "Jinkui Ren", "Xiantao Zhang", "Wen Yang", "Zhi Jin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e415074c6a4b4995350d6936d338682b9a4469f1cbed3a83e0ac44eee6f02bef", "submitted_at": "2025-09-19T02:34:50Z", "updated_at": "2025-09-19T02:34:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.15532v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.15532", "identifiers": {"arxiv": "2509.15532"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.15532"}], "name": "GUI-ARP: Enhancing Grounding with Adaptive Region Perception for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-ARP uses visual attention to crop relevant screenshot regions and decides whether a target needs single-stage or multi-stage analysis. Supervised fine-tuning followed by GRPO trains this adaptive procedure for high-resolution GUI grounding.", "topics": ["grounding", "reinforcement-learning", "screen-understanding"], "updated_at": "2025-09-19", "url": "https://arxiv.org/abs/2509.15532"}
{"aliases": [], "authors": ["Shaojie Zhang", "Ruoceng Zhang", "Pei Fu", "Shaokang Wang", "Jiahui Yang", "Xin Du", "Shiqi Cui", "Bin Qin", "Ying Huang", "Zhenbo Luo", "Jian Luan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d190eefae9309d87f4ea4780d3483ae0ebd834e0bf70e7f2af204c3788d87e7e", "submitted_at": "2025-09-19T04:03:44Z", "updated_at": "2025-10-27T11:34:58Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2509.15566v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.15566", "identifiers": {"arxiv": "2509.15566"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.15566"}], "name": "BTL-UI: Blink-Think-Link Reasoning Model for GUI Agent", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "BTL-UI separates GUI interaction into screen-region attention, planning and executable-command generation. It adds an automated annotation pipeline and rule-based process/outcome rewards for training, with evaluation on both static understanding and dynamic interaction.", "topics": ["grounding", "planning", "data-generation", "reinforcement-learning"], "updated_at": "2025-10-27", "url": "https://arxiv.org/abs/2509.15566"}
{"aliases": [], "authors": ["Musen Lin", "Minghao Liu", "Taoran Lu", "Lichen Yuan", "Yiwei Liu", "Haonan Xu", "Yu Miao", "Yuhao Chao", "Zhaojian Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "035048459d2b303721ef0738212829d1fa3314ccd0736b6649cc786c6e975ad4", "submitted_at": "2025-09-19T08:09:18Z", "updated_at": "2025-09-19T08:09:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.15738v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.15738", "identifiers": {"arxiv": "2509.15738"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.15738"}], "name": "GUI-ReWalk: Massive Data Generation for GUI Agent via Stochastic Exploration and Intent-Aware Reasoning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-ReWalk begins with stochastic interface exploration and then infers goals to generate coherent action trajectories. It supports long-horizon workflows across applications and studies training on the resulting data using grounding and navigation benchmarks.", "topics": ["data-generation", "cross-app", "long-horizon", "planning"], "updated_at": "2025-09-19", "url": "https://arxiv.org/abs/2509.15738"}
{"aliases": [], "authors": ["Hongxin Li", "Jingran Su", "Jingfan Chen", "Zheng Ju", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97b7c0690ba7deb7dbbdac6ac94dd82e3864c440cb4325bf143ad735919bb08b", "submitted_at": "2025-09-22T03:04:53Z", "updated_at": "2025-09-22T03:04:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.17328v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.17328", "identifiers": {"arxiv": "2509.17328"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.17328"}], "name": "UIPro: Unleashing Superior Interaction Capability For GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "UIPro pretrains GUI understanding on 20.6 million tasks and then harmonizes interaction datasets through a unified action space. Continued fine-tuning uses the merged multi-platform data to train action prediction for generalist GUI operation.", "topics": ["data-generation", "action-prediction", "grounding", "imitation-learning"], "updated_at": "2025-09-22", "url": "https://arxiv.org/abs/2509.17328"}
{"aliases": [], "authors": ["Tianyu Fu", "Anyang Su", "Chenxu Zhao", "Hanning Wang", "Minghui Wu", "Zhe Yu", "Fei Hu", "Mingjia Shi", "Wei Dong", "Jiayao Wang", "Yuyang Chen", "Ruiyang Yu", "Siran Peng", "Menglin Li", "Nan Huang", "Haitian Wei", "Jiawei Yu", "Yi Xin", "Xilin Zhao", "Kai Gu", "Ping Jiang", "Sifan Zhou", "Shuo Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2cc02862104494a2ecef577a0b466cb480c57ee43c611993e5d1c96716ab973d", "submitted_at": "2025-09-22T03:13:58Z", "updated_at": "2025-10-31T09:42:28Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.17336v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.17336", "identifiers": {"arxiv": "2509.17336"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.17336"}], "name": "Mano Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-31", "url": "https://arxiv.org/abs/2509.17336"}
{"aliases": [], "authors": ["Junyu Lu", "Songxin Zhang", "Zejian Xie", "Zhuoyang Song", "Jiaxing Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b85c9bae829bad343dd6f57c866ea100859fa0ad67f5c1a3ac9b3e53d617486", "submitted_at": "2025-09-22T15:40:31Z", "updated_at": "2025-09-22T15:40:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.17917v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.17917", "identifiers": {"arxiv": "2509.17917"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.17917"}], "name": "Orcust: Stepwise-Feedback Reinforcement Learning for GUI Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Orcust combines principle-constrained reward modeling with GUI trajectory collection in instrumented virtual machines. Environment-checkable and model-derived principles guide stepwise feedback for grounding, basic operations, and complete task execution.", "topics": ["reinforcement-learning", "reward-models", "data-generation", "sandbox"], "updated_at": "2025-09-22", "url": "https://arxiv.org/abs/2509.17917"}
{"aliases": [], "authors": ["Yifan Xu", "Xiao Liu", "Xinghan Liu", "Jiaqi Fu", "Hanchen Zhang", "Bohao Jing", "Shudan Zhang", "Yuting Wang", "Wenyi Zhao", "Yuxiao Dong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86bce7c1a120e0e953a239f2a9b39a04a844e6f9710525245d643932ec3c3233", "submitted_at": "2025-09-10T13:09:27Z", "updated_at": "2025-10-24T17:13:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.18119v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.18119", "identifiers": {"arxiv": "2509.18119"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.18119"}], "name": "MobileRL: Online Agentic Reinforcement Learning for Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileRL adapts GRPO to mobile GUI training with difficulty-dependent replay of successes and filtering of failure curricula. A shortest-path reward adjustment accounts for task length, with evaluations on AndroidWorld and AndroidLab.", "topics": ["reinforcement-learning", "mobile-navigation", "efficiency"], "updated_at": "2025-10-24", "url": "https://arxiv.org/abs/2509.18119"}
{"aliases": [], "authors": ["Mohammed Mehedi Hasan", "Hao Li", "Emad Fallahzadeh", "Gopi Krishnan Rajbahadur", "Bram Adams", "Ahmed E. Hassan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3dcc52aca325dbcb725946c7508958328ff77ea3c97814c67bc858ead8e27eca", "submitted_at": "2025-09-23T16:02:09Z", "updated_at": "2026-04-02T22:22:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.19185v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.19185", "identifiers": {"arxiv": "2509.19185"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.19185"}], "name": "An Empirical Study of Testing Practices in Open Source AI Agent Frameworks and Agentic Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-02", "url": "https://arxiv.org/abs/2509.19185"}
{"aliases": [], "authors": ["Xiaoqian Liu", "Ke Wang", "Yuchuan Wu", "Fei Huang", "Yongbin Li", "Junge Zhang", "Jianbin Jiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e305addce9afbc29d17a0a5ffe2cf8968286ed9102a8975d4d0cdf348d353327", "submitted_at": "2025-09-23T16:15:42Z", "updated_at": "2025-09-28T06:47:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.19199v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.19199", "identifiers": {"arxiv": "2509.19199"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.19199"}], "name": "Agentic Reinforcement Learning with Implicit Step Rewards", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-28", "url": "https://arxiv.org/abs/2509.19199"}
{"aliases": [], "authors": ["Jiazheng Sun", "Ruimeng Yang", "Xu Han", "Jiayang Niu", "Mingxuan Li", "Te Yang", "Yongyong Lu", "Xin Peng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "571eee6476c9e75ffae65cf98ec589e98fd3b953313cc6d2191bb48bd27303fe", "submitted_at": "2025-09-25T04:21:31Z", "updated_at": "2025-12-01T08:01:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.20729v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.20729", "identifiers": {"arxiv": "2509.20729"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.20729"}], "name": "Robust, Observable, and Evolvable Agentic Systems Engineering: A Principled Framework Validated via the Fairy GUI Agent", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-09-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents an agent-engineering framework combining runtime goal clarification, an observable component architecture and evolving execution memory. The Fairy mobile GUI agent evaluates these mechanisms on ambiguous tasks and through controlled and user studies.", "topics": ["observability", "human-in-the-loop", "memory", "robustness"], "updated_at": "2025-12-01", "url": "https://arxiv.org/abs/2509.20729"}
{"aliases": [], "authors": ["Kaiwen He", "Zhiwei Wang", "Chenyi Zhuang", "Jinjie Gu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db39ffa45c147739a737db392b25b1e731e7f4355e993d3df31a227f370323bb", "submitted_at": "2025-09-25T12:23:49Z", "updated_at": "2025-09-25T12:23:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.21072v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.21072", "identifiers": {"arxiv": "2509.21072"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.21072"}], "name": "Recon-Act: A Self-Evolving Multi-Agent Browser-Use System via Web Reconnaissance, Tool Generation, and Task Execution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-25", "url": "https://arxiv.org/abs/2509.21072"}
{"aliases": [], "authors": ["Yidan Zhang", "Mutian Xu", "Yiming Hao", "Kun Zhou", "Jiahao Chang", "Xiaoqiang Liu", "Pengfei Wan", "Hongbo Fu", "Xiaoguang Han"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "04b00f5f4b73d84c92c18365849a027dae0e21865b7d883b3523a823eafe4aeb", "submitted_at": "2025-09-25T15:08:28Z", "updated_at": "2025-09-25T15:08:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.21291v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.21291", "identifiers": {"arxiv": "2509.21291"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.21291"}], "name": "VC-Agent: An Interactive Agent for Customized Video Dataset Collection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-25", "url": "https://arxiv.org/abs/2509.21291"}
{"aliases": [], "authors": ["Yu Zhao", "Wei-Ning Chen", "Huseyin Atahan Inan", "Samuel Kessler", "Lu Wang", "Lukas Wutschitz", "Fangkai Yang", "Chaoyun Zhang", "Pasquale Minervini", "Saravan Rajmohan", "Robert Sim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "83fa51d3642fa7786f7f6c44fcf5d5ab5876d7ae26e5f6184ddff4d5e2009181", "submitted_at": "2025-09-25T20:38:01Z", "updated_at": "2026-05-25T20:58:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21552v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.21552", "identifiers": {"arxiv": "2509.21552"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.21552"}], "name": "Learning GUI Grounding with Spatial Reasoning from Visual Feedback", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Cursor treats GUI grounding as a sequence of cursor movements guided by visible feedback and movement history. Online reinforcement learning with trajectory rewards trains the model to refine the target location across multiple steps.", "topics": ["grounding", "reinforcement-learning", "action-prediction", "robustness"], "updated_at": "2026-05-25", "url": "https://arxiv.org/abs/2509.21552"}
{"aliases": [], "authors": ["Haotian Luo", "Huaisong Zhang", "Xuelin Zhang", "Haoyu Wang", "Zeyu Qin", "Wenjie Lu", "Guozheng Ma", "Haiying He", "Yingsha Xie", "Qiyang Zhou", "Zixuan Hu", "Hongze Mi", "Yibo Wang", "Naiqiang Tan", "Hong Chen", "Yi R. Fung", "Chun Yuan", "Li Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea80884042520482e2a0d897d121bb2c192f109792d9130957f99209512ea073", "submitted_at": "2025-09-26T02:04:00Z", "updated_at": "2025-09-26T02:04:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.21766v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.21766", "identifiers": {"arxiv": "2509.21766"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.21766"}], "name": "UltraHorizon: Benchmarking Agent Capabilities in Ultra Long-Horizon Scenarios", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-26", "url": "https://arxiv.org/abs/2509.21766"}
{"aliases": [], "authors": ["Junliang Liu", "Jingyu Xiao", "Wenxin Tang", "Zhixian Wang", "Zipeng Xie", "Wenxuan Wang", "Minrui Zhang", "Shuanghe Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8fed6f081e0330c54a304c79e8a4183865aae48f0e944545029d3de9852f8028", "submitted_at": "2025-09-26T02:38:14Z", "updated_at": "2026-03-04T16:31:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21782v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2509.21782", "identifiers": {"arxiv": "2509.21782"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.21782"}], "name": "Benchmarking MLLM-based Web Understanding: Reasoning, Robustness and Safety", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates multimodal LLMs on web interface reasoning, visual perturbation robustness, and safety-critical web understanding tasks.", "topics": ["evaluation", "screen-understanding", "safety-security", "robustness"], "updated_at": "2026-03-04", "url": "https://arxiv.org/abs/2509.21782"}
{"aliases": [], "authors": ["Hongze Mi", "Yibo Feng", "Wenjie Lu", "Yuqi Wang", "Jinyuan Li", "Song Cao", "He Cui", "Tengfei Tian", "Xuelin Zhang", "Haotian Luo", "Di Sun", "Jun Fang", "Hua Chai", "Naiqiang Tan", "Gang Pan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df0c5cb2e891375c3e6dd8e0dbebd6a41273593a3186e9eda9b1d99b05489001", "submitted_at": "2025-09-26T02:56:19Z", "updated_at": "2026-01-07T02:57:38Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.21799v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.21799", "identifiers": {"arxiv": "2509.21799"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.21799"}], "name": "D-Artemis: A Deliberative Cognitive Framework for Mobile GUI Multi-Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-07", "url": "https://arxiv.org/abs/2509.21799"}
{"aliases": [], "authors": ["Gaole Dai", "Shiqi Jiang", "Ting Cao", "Yuqing Yang", "Yuanchun Li", "Rui Tan", "Mo Li", "Lili Qiu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c96f310fb0f28888c59dce71d08697b603e7a7562fdf99b51e3504b93e9e81b3", "submitted_at": "2025-09-26T03:29:36Z", "updated_at": "2026-04-15T09:19:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21823v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.21823", "identifiers": {"arxiv": "2509.21823"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.21823"}], "name": "ProRe: A Proactive Reward System for GUI Agents via Reasoner-Actor Collaboration", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "ProRe evaluates GUI task outcomes by having a reasoner assign targeted environment-probing tasks to specialist actors. The additional observations support reward judgments, with separate evaluation of judge accuracy and downstream agent task success.", "topics": ["reward-models", "evaluation", "planning"], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2509.21823"}
{"aliases": [], "authors": ["Renqi Chen", "Zeyin Tao", "Jianming Guo", "Jingzhe Zhu", "Yiheng Peng", "Qingqing Sun", "Tianyi Zhang", "Shuai Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d23b9f33282f77b9d4965078ca993bcf21163dc1d3210aaa4e554d4fc12b37c5", "submitted_at": "2025-09-26T07:05:01Z", "updated_at": "2026-04-13T03:14:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21982v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.21982", "identifiers": {"arxiv": "2509.21982"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.21982"}], "name": "RISK: A Framework for GUI Agents in E-commerce Risk Management", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "RISK combines browser-interaction training data, a benchmark, and reinforcement fine-tuning for e-commerce risk-management workflows. Its trajectories and rewards distinguish individual actions from multi-step stateful tasks involving deeply embedded web content.", "topics": ["industry", "data-generation", "evaluation", "reinforcement-learning"], "updated_at": "2026-04-13", "url": "https://arxiv.org/abs/2509.21982"}
{"aliases": [], "authors": ["Haochen Gong", "Chenxiao Li", "Rui Chang", "Wenbo Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "922c1d9b60db7c17b9019a320a4cfc2dd4eac325fd946f024ca82d43785f5eba", "submitted_at": "2025-09-26T12:19:27Z", "updated_at": "2026-01-14T06:07:30Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2509.22256v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.22256", "identifiers": {"arxiv": "2509.22256"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.22256"}], "name": "Secure and Efficient Access Control for Computer-Use Agents via Context Space", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CSAgent, which combines static access policies with user intent and context and enforces them through an operating-system service. A supporting toolchain constructs and refines policies for agents using API, CLI and GUI interfaces.", "topics": ["safety-security", "tool-integration"], "updated_at": "2026-01-14", "url": "https://arxiv.org/abs/2509.22256"}
{"aliases": [], "authors": ["Jiho Park", "Jongyoon Song", "Minjin Choi", "Kyuho Heo", "Taehun Huh", "Ji Won Kim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce6f90688112c42387eeed2e6a8e107e5026a1f8ab2a5d8814052eeabb282892", "submitted_at": "2025-09-24T08:05:32Z", "updated_at": "2025-09-24T08:05:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.22715v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.22715", "identifiers": {"arxiv": "2509.22715"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.22715"}], "name": "TRUEBench: Can LLM Response Meet Real-world Constraints as Productivity Assistant?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-24", "url": "https://arxiv.org/abs/2509.22715"}
{"aliases": [], "authors": ["Hwan Chang", "Yonghyun Jun", "Hwanhee Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "32e4fbedf2d835af39dc005f22e530a6fcde2a5980c3f0eed28b34e7b974a6bd", "submitted_at": "2025-09-26T18:38:07Z", "updated_at": "2026-04-13T15:36:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.22830v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.22830", "identifiers": {"arxiv": "2509.22830"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.22830"}], "name": "ChatInject: Abusing Chat Templates for Prompt Injection in LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-13", "url": "https://arxiv.org/abs/2509.22830"}
{"aliases": [], "authors": ["Tao Xiong", "Xavier Hu", "Yurun Chen", "Yuhang Liu", "Changqiao Wu", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5b5375e77f087367383bb5de96cec46d8de6550d9a44bff6965b9201b6718a05", "submitted_at": "2025-09-27T11:42:36Z", "updated_at": "2026-08-30T18:12:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.23263v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2509.23263", "identifiers": {"arxiv": "2509.23263"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.23263"}], "name": "GUI-PRA: Process Reward Agent for GUI Tasks", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-PRA evaluates GUI progress by generating task-specific criteria and actively inspecting visual evidence with tools. It tests offline judgment on OS-Critic Bench and the effects of this feedback on online mobile policies in AndroidWorld.", "topics": ["reward-models", "evaluation", "screen-understanding"], "updated_at": "2026-08-30", "url": "https://arxiv.org/abs/2509.23263"}
{"aliases": [], "authors": ["Jianshuo Dong", "Sheng Guo", "Hao Wang", "Xun Chen", "Zhuotao Liu", "Tianwei Zhang", "Ke Xu", "Minlie Huang", "Han Qiu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4086e28c460007ea347c4e4c0bae15b4d3484f835763cd863d14f87fb6dda773", "submitted_at": "2025-09-28T07:05:17Z", "updated_at": "2026-06-03T04:13:57Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2509.23694v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2509.23694", "identifiers": {"arxiv": "2509.23694"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.23694"}], "name": "SafeSearch: Automated Red-Teaming of LLM-Based Search Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-03", "url": "https://arxiv.org/abs/2509.23694"}
{"aliases": [], "authors": ["Cong Chen", "Kaixiang Ji", "Hao Zhong", "Muzhi Zhu", "Anzhou Li", "Guo Gan", "Ziyuan Huang", "Cheng Zou", "Jiajia Liu", "Jingdong Chen", "Hao Chen", "Chunhua Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a2a14df6cf26113f64551d33153653891201ded849977f073408a26e1c16598", "submitted_at": "2025-09-28T08:35:16Z", "updated_at": "2025-09-28T08:35:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.23738v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.23738", "identifiers": {"arxiv": "2509.23738"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.23738"}], "name": "GUI-Shepherd: Reliable Process Reward and Verification for Long-Sequence GUI Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-28", "url": "https://arxiv.org/abs/2509.23738"}
{"aliases": [], "authors": ["Pengxiang Li", "Zechen Hu", "Zirui Shang", "Jingrong Wu", "Yang Liu", "Hui Liu", "Zhi Gao", "Chenrui Shi", "Bofei Zhang", "Zihao Zhang", "Xiaochuan Shi", "Zedong YU", "Yuwei Wu", "Xinxiao Wu", "Yunde Jia", "Liuyu Xiang", "Zhaofeng He", "Qing Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a96e774920489ef366ad462972a938863a3e714672e9c1f4b784d57eee34c4ca", "submitted_at": "2025-09-28T13:19:20Z", "updated_at": "2025-09-28T13:19:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.23866v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.23866", "identifiers": {"arxiv": "2509.23866"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.23866"}], "name": "Efficient Multi-turn RL for GUI Agents via Decoupled Training and Adaptive Data Curation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "DART separates GUI reinforcement learning into asynchronous environment, rollout, data-management, and training services. It curates experience by task difficulty, successful trajectories, and step uncertainty while correcting rollout-policy mismatch, with evaluation on OSWorld.", "topics": ["reinforcement-learning", "efficiency", "data-generation", "sandbox"], "updated_at": "2025-09-28", "url": "https://arxiv.org/abs/2509.23866"}
{"aliases": [], "authors": ["Zhecheng Li", "Guoxian Song", "Yiwei Wang", "Zhen Xiong", "Junsong Yuan", "Yujun Cai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b42e784de8dd007c60d234b7331bd98b5a04c58c13696bf436c81c7eccee37f", "submitted_at": "2025-09-29T00:06:31Z", "updated_at": "2025-09-29T00:06:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.24133v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.24133", "identifiers": {"arxiv": "2509.24133"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.24133"}], "name": "Generalist Scanner Meets Specialist Locator: A Synergistic Coarse-to-Fine Framework for Robust GUI Grounding", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "GMS pairs a general vision-language scanner that proposes interface regions with a specialist locator that predicts coordinates within them. Hierarchical coarse-to-fine search coordinates the two components for GUI grounding on ScreenSpot-Pro.", "topics": ["grounding", "planning", "evaluation"], "updated_at": "2025-09-29", "url": "https://arxiv.org/abs/2509.24133"}
{"aliases": [], "authors": ["Ran Xu", "Kaixin Ma", "Wenhao Yu", "Hongming Zhang", "Joyce C. Ho", "Carl Yang", "Dong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1c6fd3e84212bac4c4c8341ffb3478c97becce08f92dac89e5de77ad42cd07d", "submitted_at": "2025-09-29T02:04:20Z", "updated_at": "2025-09-29T02:04:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.24183v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:20:58+00:00", "role": "primary", "sha256": "fe9f70994c119a85cb8075461174e9699c423d79094e611f1ac95a51666600c1", "supports": ["summary", "environments", "scope"], "url": "https://arxiv.org/html/2509.24183v1"}], "id": "arxiv:2509.24183", "identifiers": {"arxiv": "2509.24183"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.24183"}], "name": "Retrieval-augmented GUI Agents with Generative Guidelines", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "RAG-GUI trains a small guidance adapter to filter and summarize retrieved web tutorials while keeping the acting agent fixed. Supervised and rejection-sampling stages prepare the adapter for inference-time assistance, tested on offline web/mobile tasks and interactive Android tasks.", "topics": ["memory", "planning", "imitation-learning", "web-navigation", "mobile-navigation"], "updated_at": "2025-09-29", "url": "https://arxiv.org/abs/2509.24183"}
{"aliases": [], "authors": ["Guibin Zhang", "Muxin Fu", "Shuicheng Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d1ebe5bb997f31f1afef35daf348dc34ca1583fff94ce14d7b6b40d6c60ad69a", "submitted_at": "2025-09-29T12:33:13Z", "updated_at": "2025-10-12T02:37:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.24704v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.24704", "identifiers": {"arxiv": "2509.24704"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.24704"}], "name": "MemGen: Weaving Generative Latent Memory for Self-Evolving Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-12", "url": "https://arxiv.org/abs/2509.24704"}
{"aliases": [], "authors": ["Ram Ramrakhya", "Andrew Szot", "Omar Attia", "Yuhao Yang", "Anh Nguyen", "Bogdan Mazoure", "Zhe Gan", "Harsh Agrawal", "Alexander Toshev"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d41da9264d664abbeacccae889f9f14f403fc4edbd7453ed4ff276689592fe68", "submitted_at": "2025-09-29T17:00:02Z", "updated_at": "2025-09-29T17:00:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.25047v1"}}, "environments": ["android", "linux", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2509.25047", "identifiers": {"arxiv": "2509.25047"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.25047"}], "name": "Scaling Synthetic Task Generation for Agents via Exploration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "AutoPlay explores application states and functions to generate tasks, then uses a model executor and verifier to construct demonstrations. It produces 20,000 tasks across 20 Android apps and 10,000 tasks across 13 Ubuntu apps for supervised and reinforcement learning of UI agents.", "topics": ["data-generation", "reinforcement-learning", "imitation-learning"], "updated_at": "2025-09-29", "url": "https://arxiv.org/abs/2509.25047"}
{"aliases": [], "authors": ["Siru Ouyang", "Jun Yan", "I-Hung Hsu", "Yanfei Chen", "Ke Jiang", "Zifeng Wang", "Rujun Han", "Long T. Le", "Samira Daruki", "Xiangru Tang", "Vishy Tirumalashetty", "George Lee", "Mahsan Rofouei", "Hangfei Lin", "Jiawei Han", "Chen-Yu Lee", "Tomas Pfister"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d984883b65895d02debd54edbd198f0aadf8bd295da869680c3ebbe08d2e9492", "submitted_at": "2025-09-29T17:51:03Z", "updated_at": "2026-03-16T20:49:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.25140v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2509.25140", "identifiers": {"arxiv": "2509.25140"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.25140"}], "name": "ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Distills generalizable reasoning strategies from both successful and failed agent trajectories into a persistent memory bank for test-time retrieval.", "topics": ["memory", "self-reflection", "web-navigation"], "updated_at": "2026-03-16", "url": "https://arxiv.org/abs/2509.25140"}
{"aliases": [], "authors": ["Tianrui Qin", "Qianben Chen", "Sinuo Wang", "He Xing", "King Zhu", "He Zhu", "Dingfeng Shi", "Xinxin Liu", "Ge Zhang", "Jiaheng Liu", "Yuchen Eleanor Jiang", "Xitong Gao", "Wangchunshu Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc9bcd3849ae08f4c8931cdf5e7a2e9bb482a29f2bab07a317bb604ca626979b", "submitted_at": "2025-09-29T17:39:30Z", "updated_at": "2025-09-29T17:39:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.25301v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.25301", "identifiers": {"arxiv": "2509.25301"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.25301"}], "name": "Flash-Searcher: Fast and Effective Web Agents via DAG-Based Parallel Execution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-29", "url": "https://arxiv.org/abs/2509.25301"}
{"aliases": [], "authors": ["Alexander Kovrigin", "Aleksandra Eliseeva", "Konstantin Grotov", "Egor Bogomolov", "Yaroslav Zharov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "590c631d4ba6af88650e0e66c9e19e646f918612a932cf4e2d0d424df62e3d7d", "submitted_at": "2025-09-29T20:03:05Z", "updated_at": "2025-09-29T20:03:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.25455v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.25455", "identifiers": {"arxiv": "2509.25455"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.25455"}], "name": "PIPer: On-Device Environment Setup via Online Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-09-29", "url": "https://arxiv.org/abs/2509.25455"}
{"aliases": [], "authors": ["Oshadha Wijerathne", "Amandi Nimasha", "Dushan Fernando", "Nisansa de Silva", "Srinath Perera"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45ee3d2ed438bbe807869b8d6058b76ef6383a16285085bba9ae73fc085d940e", "submitted_at": "2025-09-30T02:47:54Z", "updated_at": "2026-03-18T06:21:30Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.25693v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2509.25693", "identifiers": {"arxiv": "2509.25693"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2509.25693"}], "name": "ScheduleMe: Multi-Agent Calendar Assistant", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-18", "url": "https://arxiv.org/abs/2509.25693"}
{"aliases": [], "authors": ["Yutong Dai", "Krithika Ramakrishnan", "Jing Gu", "Matthew Fernandez", "Yanqi Luo", "Viraj Prabhu", "Zhenyu Hu", "Silvio Savarese", "Caiming Xiong", "Zeyuan Chen", "Ran Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e22d26dfbda8530f32d4686f26d91ea1a8e8c5470e857f2e13803258a2f7aaf", "submitted_at": "2025-09-30T16:48:49Z", "updated_at": "2025-09-30T16:48:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.26506v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.26506", "identifiers": {"arxiv": "2509.26506"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.26506"}], "name": "SCUBA: Salesforce Computer Use Benchmark", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "SCUBA evaluates enterprise UI navigation and CRM workflows with 300 task instances derived from Salesforce user interviews. Salesforce sandbox execution and milestone-based assessment support comparisons between zero-shot agents and agents given demonstrations.", "topics": ["evaluation", "industry", "workflow", "sandbox"], "updated_at": "2025-09-30", "url": "https://arxiv.org/abs/2509.26506"}
{"aliases": [], "authors": ["Zhen Yang", "Zi-Yi Dou", "Di Feng", "Forrest Huang", "Anh Nguyen", "Keen You", "Omar Attia", "Yuhao Yang", "Michael Feng", "Haotian Zhang", "Ram Ramrakhya", "Chao Jia", "Jeffrey Nichols", "Alexander Toshev", "Yinfei Yang", "Zhe Gan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "daa35e6bcd86fbdb709652d3592776ed93aebe2a0ebb3595fed91478dc0e3e0a", "submitted_at": "2025-09-30T17:13:56Z", "updated_at": "2025-09-30T17:13:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.26539v1"}}, "environments": ["android", "web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2509.26539", "identifiers": {"arxiv": "2509.26539"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.26539"}], "name": "Ferret-UI Lite: Lessons from Building Small On-Device GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-09-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Ferret-UI Lite develops a 3B GUI agent using mixed real and synthetic data, reasoning with visual tools at inference, and reinforcement learning. It evaluates grounding and navigation across mobile, web, and desktop settings, including AndroidWorld and OSWorld.", "topics": ["grounding", "reinforcement-learning", "tool-integration", "efficiency"], "updated_at": "2025-09-30", "url": "https://arxiv.org/abs/2509.26539"}
{"aliases": [], "authors": ["Amirhossein Abaskohi", "Tianyi Chen", "Miguel Muñoz-Mármol", "Curtis Fox", "Amrutha Varshini Ramesh", "Étienne Marcotte", "Xing Han Lù", "Nicolas Chapados", "Spandana Gella", "Peter West", "Giuseppe Carenini", "Christopher Pal", "Alexandre Drouin", "Issam H. Laradji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4d8ff34091c0515b3a239380bdd20cdda00f8d779cf0979b77439b4fef4fc3f4", "submitted_at": "2025-09-30T18:47:20Z", "updated_at": "2026-03-10T00:07:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.00172v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.00172", "identifiers": {"arxiv": "2510.00172"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.00172"}], "name": "DRBench: A Realistic Benchmark for Enterprise Deep Research", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-10", "url": "https://arxiv.org/abs/2510.00172"}
{"aliases": [], "authors": ["Zikang Liu", "Junyi Li", "Wayne Xin Zhao", "Dawei Gao", "Yaliang Li", "Ji-rong Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "277315ecc1a341c0a511bce84b87190fc0aeeb2306666ea8d4746aa8e9b39984", "submitted_at": "2025-10-01T01:48:39Z", "updated_at": "2025-10-04T15:23:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.00413v2"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.00413", "identifiers": {"arxiv": "2510.00413"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.00413"}], "name": "PAL-UI: Planning with Active Look-back for Vision-Based GUI Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "PAL-UI combines observation/action summaries with selective retrieval of historical screenshots during planning. It trains mobile navigation models from step-level supervision and evaluates transfer to web navigation without additional training.", "topics": ["memory", "planning", "interaction-history", "mobile-navigation"], "updated_at": "2025-10-04", "url": "https://arxiv.org/abs/2510.00413"}
{"aliases": [], "authors": ["Zichen Wen", "Shaobo Wang", "Yufa Zhou", "Junyuan Zhang", "Qintong Zhang", "Yifeng Gao", "Zhaorun Chen", "Bin Wang", "Weijia Li", "Conghui He", "Linfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "59ab271123e55be633f79291d2846cfdafbf7536d6802aec4e280404df7ff6a2", "submitted_at": "2025-10-01T04:56:40Z", "updated_at": "2025-10-01T04:56:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.00515v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.00515", "identifiers": {"arxiv": "2510.00515"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.00515"}], "name": "Efficient Multi-modal Large Language Models via Progressive Consistency Distillation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-01", "url": "https://arxiv.org/abs/2510.00515"}
{"aliases": [], "authors": ["Kung-Hsiang Huang", "Haoyi Qiu", "Yutong Dai", "Caiming Xiong", "Chien-Sheng Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e07a5a8f9c0aa39643bb919c2d637435669318420223d7d72b0ef3d733308852", "submitted_at": "2025-10-01T05:37:54Z", "updated_at": "2026-10-07T23:23:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.00536v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.00536", "identifiers": {"arxiv": "2510.00536"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.00536"}], "name": "GUI-KV: Efficient GUI Agents via KV Cache with Spatio-Temporal Awareness", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-KV compresses vision-language agents’ attention caches using visual-token saliency and redundancy between screenshot frames. The method requires no retraining and evaluates how cache budgets affect GUI action accuracy and computational cost.", "topics": ["efficiency", "interaction-history", "training-free", "screen-understanding"], "updated_at": "2026-10-07", "url": "https://arxiv.org/abs/2510.00536"}
{"aliases": [], "authors": ["Yinuo Liu", "Xilong Wang", "Ruohan Xu", "Yuqi Jia", "Neil Zhenqiang Gong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e997e87268cfe66869bf19fdc8445e5487c980736a63cbe76f779bd1f142c396", "submitted_at": "2025-10-01T18:34:06Z", "updated_at": "2026-09-23T16:44:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.01354v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.01354", "identifiers": {"arxiv": "2510.01354"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.01354"}], "name": "WAInjectBench: Benchmarking Prompt Injection Detections for Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The reported contribution evaluates attack detection; end-to-end prevention of browser-agent compromise is not inferred.", "published_at": "2025-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WAInjectBench with a threat-model taxonomy and malicious and benign text-image datasets for web-agent prompt-injection detection. It compares detector families across explicit and less visible forms of injected content.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-09-23", "url": "https://arxiv.org/abs/2510.01354"}
{"aliases": [], "authors": ["Viraj Prabhu", "Yutong Dai", "Matthew Fernandez", "Jing Gu", "Krithika Ramakrishnan", "Yanqi Luo", "Silvio Savarese", "Caiming Xiong", "Junnan Li", "Zeyuan Chen", "Ran Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "23794419031fe219fc342c041d666045080b274926dc22ab101a65581b4e4efd", "submitted_at": "2025-10-01T23:41:47Z", "updated_at": "2025-10-01T23:41:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.01524v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:27:51+00:00", "role": "primary", "sha256": "8466eed10f7f630a68fa742ff9de6b5ba052c497ff421acd4fbee529093d3a58", "supports": ["summary", "scope", "environments"], "url": "https://arxiv.org/html/2510.01524v1"}], "id": "arxiv:2510.01524", "identifiers": {"arxiv": "2510.01524"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.01524"}], "name": "WALT: Web Agents that Learn Tools", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "WALT explores website functionality and converts demonstrated interactions into validated, reusable tools combining DOM actions, URL navigation, and agentic fallbacks. It evaluates this expanded browser action space on WebArena and VisualWebArena.", "topics": ["web-navigation", "tool-integration", "efficiency", "automation"], "updated_at": "2025-10-01", "url": "https://arxiv.org/abs/2510.01524"}
{"aliases": [], "authors": ["Erfan Shayegani", "Keegan Hines", "Yue Dong", "Nael Abu-Ghazaleh", "Roman Lutz", "Spencer Whitehead", "Vidhisha Balachandran", "Besmira Nushi", "Vibhav Vineet"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a6426fa5eeb3dfafef30e47dd46a6a43dff3e24bf0ff485b7d10ab9efe133f94", "submitted_at": "2025-10-02T04:52:15Z", "updated_at": "2025-10-02T04:52:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.01670v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.01670", "identifiers": {"arxiv": "2510.01670"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.01670"}], "name": "Just Do It!? Computer-Use Agents Exhibit Blind Goal-Directedness", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "BLIND-ACT tests whether computer-use agents persist with ambiguous, contradictory, or infeasible goals without adequate contextual reasoning. Its OSWorld-based tasks and judged trajectories examine these behaviors and the effect of mitigation prompts.", "topics": ["evaluation", "safety-security", "robustness"], "updated_at": "2025-10-02", "url": "https://arxiv.org/abs/2510.01670"}
{"aliases": [], "authors": ["Yang Yao", "Yixu Wang", "Yuxuan Zhang", "Yi Lu", "Tianle Gu", "Lingyu Li", "Dingyi Zhao", "Keming Wu", "Haozhe Wang", "Ping Nie", "Yan Teng", "Yingchun Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3fce66bf32563ba2fb5fd6d9d04d751837216e61eb1bf4d65deee53e7263dd6e", "submitted_at": "2025-10-02T16:40:02Z", "updated_at": "2026-01-29T14:47:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.02190v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.02190", "identifiers": {"arxiv": "2510.02190"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.02190"}], "name": "Dr. Bench: A Multidimensional Evaluation for Deep Research Agents, from Answers to Reports", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-29", "url": "https://arxiv.org/abs/2510.02190"}
{"aliases": [], "authors": ["Lingzhong Dong", "Ziqi Zhou", "Shuaibo Yang", "Haiyue Sheng", "Pengzhou Cheng", "Zongru Wu", "Zheng Wu", "Gongshen Liu", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cef1e2bf8449f4a79585492074cc80536b0af7c6e4d87c823de7005f888ec377", "submitted_at": "2025-10-02T16:51:19Z", "updated_at": "2025-10-02T16:51:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.02204v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.02204", "identifiers": {"arxiv": "2510.02204"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.02204"}], "name": "Say One Thing, Do Another? Diagnosing Reasoning-Execution Gaps in VLM-Powered Mobile-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-02", "url": "https://arxiv.org/abs/2510.02204"}
{"aliases": [], "authors": ["Gonzalo Gonzalez-Pumariega", "Vincent Tu", "Chih-Lun Lee", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29c00ba4713da26f912fec8046d498b4fd10cf71a3109f66e92fa23f7cebcf44", "submitted_at": "2025-10-02T17:37:08Z", "updated_at": "2026-02-03T23:48:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.02250v2"}}, "environments": ["desktop", "windows", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.02250", "identifiers": {"arxiv": "2510.02250"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.02250"}], "name": "Scaling Agents for Computer Use", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Behavior Judge, which converts execution trajectories into behavior narratives and compares multiple candidate rollouts. Experiments examine this selection approach on OSWorld, WindowsAgentArena and AndroidWorld.", "topics": ["evaluation", "reward-models", "long-horizon"], "updated_at": "2026-02-03", "url": "https://arxiv.org/abs/2510.02250"}
{"aliases": [], "authors": ["Sagnik Anupam", "Davis Brown", "Shuo Li", "Eric Wong", "Hamed Hassani", "Osbert Bastani"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7906e06c2efb8778a43ec774821f90947e70cec0b78d6d43ac8901f7f0562332", "submitted_at": "2025-10-02T15:22:21Z", "updated_at": "2025-10-07T15:12:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.02418v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.02418", "identifiers": {"arxiv": "2510.02418"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.02418"}], "name": "BrowserArena: Evaluating LLM Agents on Real-World Web Navigation Tasks", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces BrowserArena, a live-web platform for head-to-head agent comparisons on user-submitted tasks. Step-level human annotations support targeted analysis of navigation, pop-up handling and CAPTCHA-related failures.", "topics": ["evaluation", "web-navigation", "human-in-the-loop"], "updated_at": "2025-10-07", "url": "https://arxiv.org/abs/2510.02418"}
{"aliases": [], "authors": ["Imene Kerboua", "Sahar Omidi Shayegan", "Megh Thakkar", "Xing Han Lù", "Léo Boisvert", "Massimo Caccia", "Jérémy Espinas", "Alexandre Aussem", "Véronique Eglin", "Alexandre Lacoste"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "be30412c1d55634c466e062de8482478e298b8b447ae5996899a8a72cf5aa23c", "submitted_at": "2025-10-03T17:41:30Z", "updated_at": "2026-08-29T16:53:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.03204v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.03204", "identifiers": {"arxiv": "2510.03204"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.03204"}], "name": "FocusAgent: Simple Yet Effective Ways of Trimming the Large Context of Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-29", "url": "https://arxiv.org/abs/2510.03204"}
{"aliases": [], "authors": ["Suyuchen Wang", "Tianyu Zhang", "Ahmed Masry", "Christopher Pal", "Spandana Gella", "Bang Liu", "Perouz Taslakian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f0c7eedadb9a321b222d4081271d6c979a25e1f64714cea8cf2ff088c5e61df", "submitted_at": "2025-10-03T17:59:34Z", "updated_at": "2025-10-03T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.03230v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.03230", "identifiers": {"arxiv": "2510.03230"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.03230"}], "name": "Improving GUI Grounding with Explicit Position-to-Coordinate Mapping", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "This work addresses GUI localization errors caused by mismatches between visual patch positions and output pixel coordinates. RULER tokens provide explicit coordinate references, while Interleaved MRoPE balances width and height encoding; evaluation uses the ScreenSpot benchmark family.", "topics": ["grounding", "screen-understanding", "robustness"], "updated_at": "2025-10-03", "url": "https://arxiv.org/abs/2510.03230"}
{"aliases": [], "authors": ["Su Kara", "Fazle Faisal", "Suman Nath"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88e25f6d876dc6ad11219e3faa0e7d95f4abb39a9aba32213896225aff64ec42", "submitted_at": "2025-09-28T20:51:05Z", "updated_at": "2025-09-28T20:51:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.03285v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.03285", "identifiers": {"arxiv": "2510.03285"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.03285"}], "name": "WAREX: Web Agent Reliability Evaluation on Existing Benchmarks", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "WAREX augments existing web-agent benchmarks with reliability challenges arising from unstable systems and changing or disruptive website behavior. Evaluation on WebArena, WebVoyager and REAL examines how these disturbances affect task completion.", "topics": ["evaluation", "robustness", "web-navigation"], "updated_at": "2025-09-28", "url": "https://arxiv.org/abs/2510.03285"}
{"aliases": [], "authors": ["Xiaofei Wen", "Wenjie Jacky Mo", "Yanan Xie", "Peng Qi", "Muhao Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d96a41d1537f8c1145d3fbcd770bff8a6f8f663df0149926ce001c7bd86ecf9", "submitted_at": "2025-10-03T20:03:19Z", "updated_at": "2026-05-18T21:29:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.03485v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2510.03485", "identifiers": {"arxiv": "2510.03485"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.03485"}], "name": "Learning Efficient Guardrails for Compliance", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-18", "url": "https://arxiv.org/abs/2510.03485"}
{"aliases": [], "authors": ["Tanqiu Jiang", "Min Bai", "Nikolaos Pappas", "Yanjun Qi", "Sandesh Swamy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1389b3ccafad820f5b9f5c306bbcdbaa27fe4d5918c83bb8535c9eadd6ff2468", "submitted_at": "2025-10-04T01:57:20Z", "updated_at": "2025-10-04T01:57:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.03612v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.03612", "identifiers": {"arxiv": "2510.03612"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.03612"}], "name": "Cross-Modal Content Optimization for Steering Web Agent Preferences", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-04", "url": "https://arxiv.org/abs/2510.03612"}
{"aliases": [], "authors": ["Bin Lei", "Nuo Xu", "Ali Payani", "Mingyi Hong", "Chunhua Liao", "Yu Cao", "Caiwen Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "798fc5ca4fe483a1235f0979b0c641e9c1732660780c13340c5221319a1b28fc", "submitted_at": "2025-10-05T05:15:45Z", "updated_at": "2025-10-05T05:15:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.04039v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.04039", "identifiers": {"arxiv": "2510.04039"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.04039"}], "name": "\\textsc{GUI-Spotlight}: Adaptive Iterative Focus Refinement for Enhanced GUI Visual Grounding", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Spotlight trains a grounding model to reason over images and call tools that iteratively narrow its visual focus. The paper evaluates this refinement method on ScreenSpot-Pro after training on 18,500 samples.", "topics": ["grounding", "planning", "tool-integration"], "updated_at": "2025-10-05", "url": "https://arxiv.org/abs/2510.04039"}
{"aliases": [], "authors": ["Hadi Nekoei", "Aman Jaiswal", "Patrice Bechard", "Oleh Shliazhko", "Orlando Marquez Ayala", "Mathieu Reymond", "Massimo Caccia", "Alexandre Drouin", "Sarath Chandar", "Alexandre Lacoste"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d4bd517ada9303e765827b3badfa2ee432bc5511ff4ee0cc42d937735e1c8de", "submitted_at": "2025-10-05T21:34:42Z", "updated_at": "2026-02-20T19:34:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.04373v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.04373", "identifiers": {"arxiv": "2510.04373"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.04373"}], "name": "JEF-Hinter: Leveraging Offline Knowledge for Improving Web Agents Adaptation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces JEF-Hinter, which distills successful and failed trajectories into compact episodic hints and retrieves them for the current agent state. Web benchmark experiments examine guidance from decisive steps rather than full demonstration replay.", "topics": ["memory", "interaction-history", "web-navigation"], "updated_at": "2026-02-20", "url": "https://arxiv.org/abs/2510.04373"}
{"aliases": [], "authors": ["Jenny T. Liang", "Titus Barik", "Jeffrey Nichols", "Eldon Schoop", "Ruijia Cheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15bc41d6223eb659fd4d99cb55b83b1dc60cdc98cf990cb641311ab1879c841f", "submitted_at": "2025-10-06T02:58:42Z", "updated_at": "2026-07-28T07:01:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.04452v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.04452", "identifiers": {"arxiv": "2510.04452"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.04452"}], "name": "Understanding User Experiences of Computer Use Agents: Design Space and Opportunities for Building Agent UX Prototypes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-28", "url": "https://arxiv.org/abs/2510.04452"}
{"aliases": [], "authors": ["Muyu He", "Anand Kumar", "Tsach Mackey", "Meghana Rajeev", "James Zou", "Nazneen Rajani"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af14c68343e8db9d4312d250c095dec27de88fcc52352b333a08854820ead392", "submitted_at": "2025-10-06T05:03:57Z", "updated_at": "2026-06-08T20:35:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.04491v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.04491", "identifiers": {"arxiv": "2510.04491"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.04491"}], "name": "Impatient Users Confuse AI Agents: High-fidelity Simulations of Human Traits for Testing Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-08", "url": "https://arxiv.org/abs/2510.04491"}
{"aliases": [], "authors": ["Yuan Wang", "Mingyu Li", "Haibo Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c22547bf82eb2fb87e8bf6a3846d25c17cac296763726ed9a25faf88bdca33c", "submitted_at": "2025-10-06T09:14:58Z", "updated_at": "2026-03-25T11:29:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.04607v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.04607", "identifiers": {"arxiv": "2510.04607"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.04607"}], "name": "From Imperative to Declarative: Towards LLM-friendly OS Interfaces for Boosted Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a Declarative Model Interface that maps existing GUIs to access, state and observation primitives while handling low-level interactions. Evaluation uses Word, PowerPoint and Excel on Windows with a GUI-agent baseline.", "topics": ["automation", "tool-integration", "planning"], "updated_at": "2026-03-25", "url": "https://arxiv.org/abs/2510.04607"}
{"aliases": [], "authors": ["Chan Hee Song", "Yiwen Song", "Palash Goyal", "Yu Su", "Oriana Riva", "Hamid Palangi", "Tomas Pfister"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f2d5507adeb42d289dd3fe654c097d7772b331bdc1754011d9a53bdaab21a60", "submitted_at": "2025-10-06T10:29:00Z", "updated_at": "2026-03-16T04:38:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.04673v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2510.04673", "identifiers": {"arxiv": "2510.04673"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.04673"}], "name": "Watch and Learn: Learning to Use Computers from Online Videos", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-16", "url": "https://arxiv.org/abs/2510.04673"}
{"aliases": [], "authors": ["Zeyu Zhu", "Kevin Qinghong Lin", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "285fbce93418796cc9a7e8452a6070af04bb603d7af07bc963e93f1d1492589e", "submitted_at": "2025-10-06T17:58:02Z", "updated_at": "2025-10-09T17:29:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.05096v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.05096", "identifiers": {"arxiv": "2510.05096"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.05096"}], "name": "Paper2Video: Automatic Video Generation from Scientific Papers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-09", "url": "https://arxiv.org/abs/2510.05096"}
{"aliases": [], "authors": ["Lesly Miculicich", "Mihir Parmar", "Hamid Palangi", "Krishnamurthy Dj Dvijotham", "Mirko Montanari", "Tomas Pfister", "Long T. Le"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7abeac61a858e549aa734c4d675b6a64adf12a3d0ad1ecafeb84ee555c3a2cd0", "submitted_at": "2025-10-03T04:11:43Z", "updated_at": "2025-10-03T04:11:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.05156v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.05156", "identifiers": {"arxiv": "2510.05156"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.05156"}], "name": "VeriGuard: Enhancing LLM Agent Safety via Verified Code Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-03", "url": "https://arxiv.org/abs/2510.05156"}
{"aliases": [], "authors": ["Nikita Pavlichenko", "Iurii Nazarov", "Ivan Dolgov", "Ekaterina Garanina", "Dmitry Ustalov", "Ivan Bondyrev", "Kseniia Lysaniuk", "Evgeniia Vu", "Kirill Chekmenev", "Joseph Shtok", "Yaroslav Golubev", "Anton Semenkin", "Uladzislau Sazanovich"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bbeceac03b5f51a3e59710c9d8d1287b217d98f084714fa42f00f3fbeefd3484", "submitted_at": "2025-10-07T11:09:11Z", "updated_at": "2025-10-07T11:09:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.05788v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.05788", "identifiers": {"arxiv": "2510.05788"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.05788"}], "name": "Mellum: Production-Grade in-IDE Contextual Code Completion with Multi-File Project Understanding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-07", "url": "https://arxiv.org/abs/2510.05788"}
{"aliases": [], "authors": ["Shang Ma", "Xusheng Xiao", "Yanfang Ye"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4bf0a561460c080b14da6fd13f6d92781e01e52b08dc691fe019e70b88e44c86", "submitted_at": "2025-10-07T15:36:04Z", "updated_at": "2026-09-27T17:25:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.06042v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:29.861666+00:00", "role": "primary", "sha256": "5e2cb2f053983140904ec15c097d060ff18315c6c3c956bbcacbec26921709c5", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2510.06042v3"}], "id": "arxiv:2510.06042", "identifiers": {"arxiv": "2510.06042"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:10.946679+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.06042"}], "name": "Agent+P: Guiding UI Agents via Symbolic Planning", "notes": "Planning guarantees are relative to the modeled graph and selected target. They do not establish complete graphs, correct grounding or faithful goal translation. The described design depends on static source analysis; no experiments were reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3.3 UI Automation to Classical Planning; 4.1 UTG Builder; 4.2 Node Selector; 4.3 Plan Generator; 4.4 UI Explorer; 5.1 Datasets); not an entire-paper reading or reproduction claim.", "published_at": "2025-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Guides UI agents with symbolic planning over a UI transition graph initialized from app-code analysis and updated through online exploration, with evaluation on AndroidWorld.", "topics": ["planning", "long-horizon", "mobile-navigation"], "updated_at": "2026-09-27", "url": "https://arxiv.org/abs/2510.06042"}
{"aliases": [], "authors": ["Hans G. W. van Dam"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9538fa696994dc6ccb25cdc6c3f07e4745063f87bbfd9250e930100092965b83", "submitted_at": "2025-08-31T14:40:11Z", "updated_at": "2025-10-09T12:55:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.06223v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.06223", "identifiers": {"arxiv": "2510.06223"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.06223"}], "name": "A Multimodal GUI Architecture for Interfacing with LLM-Based Conversational Assistants", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-08-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-09", "url": "https://arxiv.org/abs/2510.06223"}
{"aliases": [], "authors": ["Weidi Luo", "Qiming Zhang", "Tianyu Lu", "Xiaogeng Liu", "Bin Hu", "Hung-Chun Chiu", "Siyuan Ma", "Yizhe Zhang", "Xusheng Xiao", "Yinzhi Cao", "Zhen Xiang", "Chaowei Xiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43be43a4abda0407e20693d43c0f6304e384d60becbcea81b4a6cdc64d852059", "submitted_at": "2025-10-08T03:35:23Z", "updated_at": "2026-08-21T05:20:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.06607v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.06607", "identifiers": {"arxiv": "2510.06607"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.06607"}], "name": "Your Harness is Not Secure: Benchmarking Real-world Threat of Command Line Interface Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-21", "url": "https://arxiv.org/abs/2510.06607"}
{"aliases": [], "authors": ["MohammadHossein Rezaei", "Robert Vacareanu", "Zihao Wang", "Clinton Wang", "Bing Liu", "Yunzhong He", "Afra Feyza Akyürek"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "997c504039e0b8d571509252da16678c361533365333e3a99b035b6e97fe3983", "submitted_at": "2025-10-08T17:44:59Z", "updated_at": "2025-10-09T18:26:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.07284v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.07284", "identifiers": {"arxiv": "2510.07284"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.07284"}], "name": "Online Rubrics Elicitation from Pairwise Comparisons", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-09", "url": "https://arxiv.org/abs/2510.07284"}
{"aliases": [], "authors": ["Debadrito Roy", "Aryaman Manish Kolhe", "V. Lalitha", "Navin Kashyap"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "729eddd4d6c75381ae286635ed8aecd97e9d92f735c377c51570adc2728e56c6", "submitted_at": "2025-10-08T20:49:49Z", "updated_at": "2025-10-10T09:52:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.07532v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.07532", "identifiers": {"arxiv": "2510.07532"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.07532"}], "name": "On Quantum Computation Using Bias-Preserving Gates", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-10", "url": "https://arxiv.org/abs/2510.07532"}
{"aliases": [], "authors": ["Renhua Ding", "Xiao Yang", "Zhengwei Fang", "Jun Luo", "Kun He", "Jun Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e80999148931b7485d51ef87887d06240f813747a9739282ed0daf53a912f9f6", "submitted_at": "2025-10-09T05:34:57Z", "updated_at": "2026-04-09T01:41:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2510.07809v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2510.07809", "identifiers": {"arxiv": "2510.07809"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.07809"}], "name": "Invisible to Humans, Triggered by Agents: Stealthy Jailbreak Attacks on Mobile Vision-Language Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses the near-zero contact-touch signals of automated mobile agents to expose malicious visual content selectively during agent interactions. The paper combines this agent-only perceptual injection with HG-IDA* prompt optimization to study unauthorized cross-app actions.", "topics": ["safety-security", "mobile-navigation", "robustness"], "updated_at": "2026-04-09", "url": "https://arxiv.org/abs/2510.07809"}
{"aliases": [], "authors": ["Peiran Wu", "Zhuorui Yu", "Yunze Liu", "Chi-Hao Wu", "Enmin Zhou", "Junxiao Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "79002b0990f4009953c8af66e55298adb6a77e134b8e1ffa84180ac78ce3e324", "submitted_at": "2025-10-09T08:07:19Z", "updated_at": "2026-02-09T10:50:38Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.07915v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.07915", "identifiers": {"arxiv": "2510.07915"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.07915"}], "name": "MARC: Memory-Augmented RL Token Compression for Efficient Video Understanding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-09", "url": "https://arxiv.org/abs/2510.07915"}
{"aliases": [], "authors": ["Haitao Jia", "Ming He", "Zimo Yin", "Likang Wu", "Jianping Fan", "Jitao Sang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9460a0926936d3db23c36c6c0e069d664265a8a9513c08c1a04f5a49970d6567", "submitted_at": "2025-10-09T09:22:05Z", "updated_at": "2025-10-09T09:22:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.07988v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.07988", "identifiers": {"arxiv": "2510.07988"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.07988"}], "name": "ReInAgent: A Context-Aware GUI Agent Enabling Human-in-the-Loop Mobile Task Navigation", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ReInAgent, coordinating information management, planning, and reflection around a shared memory for mobile tasks. It proactively asks users for information and checks consistency when instructions are ambiguous or change during execution.", "topics": ["human-in-the-loop", "memory", "mobile-navigation"], "updated_at": "2025-10-09", "url": "https://arxiv.org/abs/2510.07988"}
{"aliases": [], "authors": ["Yiming Li", "Nael Darwiche", "Amirreza Razmjoo", "Sichao Liu", "Yilun Du", "Auke Ijspeert", "Sylvain Calinon"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fda97c393b8e378b3bcef8db7abf387b1c5e58528e8edafe38afb1e940599fc9", "submitted_at": "2025-10-09T20:09:51Z", "updated_at": "2025-10-09T20:09:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.08787v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.08787", "identifiers": {"arxiv": "2510.08787"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.08787"}], "name": "Geometry-aware Policy Imitation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-09", "url": "https://arxiv.org/abs/2510.08787"}
{"aliases": [], "authors": ["Wenyi Wu", "Kun Zhou", "Ruoxin Yuan", "Vivian Yu", "Stephen Wang", "Zhiting Hu", "Biwei Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43c41dfbef8b76bbdccd7c435694cee8788b116db278122f64a22e7498bd5a5a", "submitted_at": "2025-10-10T06:16:45Z", "updated_at": "2025-10-10T06:16:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.09038v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.09038", "identifiers": {"arxiv": "2510.09038"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.09038"}], "name": "Auto-scaling Continuous Memory for GUI Agent", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-10-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Encodes GUI trajectories into compact continuous embeddings that can be retrieved directly into an agent’s input. An automated environment-discovery and rollout pipeline expands this memory for long tasks and unfamiliar interfaces.", "topics": ["memory", "interaction-history", "efficiency"], "updated_at": "2025-10-10", "url": "https://arxiv.org/abs/2510.09038"}
{"aliases": [], "authors": ["Enze Zhang", "Jiaying Wang", "Mengxi Xiao", "Jifei Liu", "Ziyan Kuang", "Rui Dong", "Eric Dong", "Sophia Ananiadou", "Min Peng", "Qianqian Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "748e1a62c4db627fcf1f747d60ac9aa59de24a5a70f14a8ea4da793c1624111e", "submitted_at": "2025-10-10T08:10:10Z", "updated_at": "2025-10-13T05:51:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.09116v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2510.09116", "identifiers": {"arxiv": "2510.09116"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.09116"}], "name": "DITING: A Multi-Agent Evaluation Framework for Benchmarking Web Novel Translation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-13", "url": "https://arxiv.org/abs/2510.09116"}
{"aliases": [], "authors": ["Yuchen Lu", "Run Yang", "Yichen Zhang", "Shuguang Yu", "Ziwei Wang", "Jiayi Xiang", "Wenxin E", "Changyu Zhu", "Fan Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d7bd622186ac5724bc7354b2bacb390c0ecb037e4a4d3182cc4d4aa3f0ea9129", "submitted_at": "2025-10-10T16:28:43Z", "updated_at": "2026-07-02T09:56:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.09517v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}], "id": "arxiv:2510.09517", "identifiers": {"arxiv": "2510.09517"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.09517"}], "name": "StatEval: A Comprehensive Benchmark for Large Language Models in Statistics", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-02", "url": "https://arxiv.org/abs/2510.09517"}
{"aliases": [], "authors": ["Sanjari Srivastava", "Gang Li", "Cheng Chang", "Rishu Garg", "Manpreet Kaur", "Charlene Y. Lee", "Yuezhang Li", "Yining Mao", "Ignacio Cases", "Yanan Xie", "Peng Qi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3af3fb75d012ebcf4ad082ab7b83b0553596bbbb8598751d0a70c86d929a22ee", "submitted_at": "2025-10-10T21:20:51Z", "updated_at": "2026-05-18T22:06:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.09872v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.09872", "identifiers": {"arxiv": "2510.09872"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.09872"}], "name": "WARC-Bench: Web Archive Based Benchmark for GUI Subtask Executions", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "WARC-Bench packages dynamic web pages as replayable archives and evaluates short navigation subtasks involving components such as date pickers and scrollable containers. It also studies supervised training followed by reinforcement learning with verifiable rewards on these browser interactions.", "topics": ["evaluation", "web-navigation", "sandbox", "reinforcement-learning"], "updated_at": "2026-05-18", "url": "https://arxiv.org/abs/2510.09872"}
{"aliases": [], "authors": ["Zonghao Ying", "Yangguang Shao", "Jianle Gan", "Gan Xu", "Wenxin Zhang", "Quanchen Zou", "Junzheng Shi", "Zhenfei Yin", "Mingchuan Zhang", "Aishan Liu", "Xianglong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43c9f7c91539df587a4f08888820731aef5bb3fca2cfb9e742443330aaf1c35f", "submitted_at": "2025-10-11T07:18:12Z", "updated_at": "2026-04-14T02:53:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.10073v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.10073", "identifiers": {"arxiv": "2510.10073"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.10073"}], "name": "SecureWebArena: A Holistic Security Evaluation Benchmark for LVLM-based Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SecureWebArena with simulated interactive websites, an attack-vector taxonomy and labeled agent trajectories. Its evaluation separates failures in reasoning, behavioral traces and task outcomes across vision-language agent families.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2510.10073"}
{"aliases": [], "authors": ["Jingchao Wang", "Wenlong Zhang", "Dingjiang Huang", "Hong Wang", "Yefeng Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c3518df14fc6d92f640dca5b20d666f8802b45ed5914d94e4a2ac4cca813321b", "submitted_at": "2025-10-12T13:06:59Z", "updated_at": "2025-10-12T13:06:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.10587v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.10587", "identifiers": {"arxiv": "2510.10587"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.10587"}], "name": "A Simple and Better Baseline for Visual Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-12", "url": "https://arxiv.org/abs/2510.10587"}
{"aliases": [], "authors": ["Tao Yu", "Zhengbo Zhang", "Zhiheng Lyu", "Junhao Gong", "Hongzhu Yi", "Xinming Wang", "Yuxuan Zhou", "Jiabing Yang", "Ping Nie", "Yan Huang", "Wenhu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe982af6e7285c21e13dcd4565255b501b401b188da64069d9cfe2fab7891f88", "submitted_at": "2025-10-12T15:43:37Z", "updated_at": "2025-10-14T08:54:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.10666v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.10666", "identifiers": {"arxiv": "2510.10666"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.10666"}], "name": "BrowserAgent: Building Web Agents with Human-Inspired Web Browsing Actions", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains BrowserAgent to use Playwright-based scrolling, clicking and typing directly on webpages for information-seeking tasks. Supervised and rejection fine-tuning are paired with explicit memory for conclusions carried across steps.", "topics": ["web-navigation", "imitation-learning", "memory"], "updated_at": "2025-10-14", "url": "https://arxiv.org/abs/2510.10666"}
{"aliases": [], "authors": ["Daoyu Wang", "Mingyue Cheng", "Shuo Yu", "Zirui Liu", "Ze Guo", "Xin Li", "Qi Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5990c9223d3391b54d9a42a74cbed99986f336016812fefdfaeccce5fb454821", "submitted_at": "2025-10-13T02:10:39Z", "updated_at": "2026-01-30T02:57:53Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2510.10909v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.10909", "identifiers": {"arxiv": "2510.10909"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.10909"}], "name": "PaperArena: An Evaluation Benchmark for Tool-Augmented Agentic Reasoning on Scientific Literature", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-30", "url": "https://arxiv.org/abs/2510.10909"}
{"aliases": [], "authors": ["Longjie Guo", "Chenjie Yuan", "Mingyuan Zhong", "Robert Wolfe", "Ruican Zhong", "Yue Xu", "Bingbing Wen", "Hua Shen", "Lucy Lu Wang", "Alexis Hiniker"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6152a61f81c435e76a432bc10086827c0cf0d2f44026bac03d6b80ebae8ca77", "submitted_at": "2025-10-13T06:12:03Z", "updated_at": "2026-02-23T08:31:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.11035v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.11035", "identifiers": {"arxiv": "2510.11035"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.11035"}], "name": "SusBench: An Online Benchmark for Evaluating Dark Pattern Susceptibility of Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SusBench, an online benchmark that inserts dark-pattern designs into consumer websites for agent evaluation. A human study checks the realism of the interventions, and agent trials compare susceptibility across pattern types.", "topics": ["ui-design", "safety-security", "evaluation"], "updated_at": "2026-02-23", "url": "https://arxiv.org/abs/2510.11035"}
{"aliases": [], "authors": ["Gautier Dagan", "Frank Keller", "Alex Lascarides"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "771718086d78de5ed3e2bf3d84aa78d4a031e7a2cdac15aeb7a4bfc3831e535d", "submitted_at": "2025-10-13T08:35:20Z", "updated_at": "2026-05-02T07:11:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.11144v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.11144", "identifiers": {"arxiv": "2510.11144"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.11144"}], "name": "$How^{2}$: How to learn from procedural How-to questions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-02", "url": "https://arxiv.org/abs/2510.11144"}
{"aliases": [], "authors": ["Tao Li", "Jinlong Hu", "Yang Wang", "Junfeng Liu", "Xuejun Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "baa5f70d4b6516b2c7fdda997ce3c4d074e26ccfd9a3ad0c00f83177bd9d4d4b", "submitted_at": "2025-10-13T10:05:43Z", "updated_at": "2025-10-13T10:05:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.11221v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.11221", "identifiers": {"arxiv": "2510.11221"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.11221"}], "name": "WebRouter: Query-specific Router via Variational Information Bottleneck for Cost-sensitive Web Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-13", "url": "https://arxiv.org/abs/2510.11221"}
{"aliases": [], "authors": ["Kai Mei", "Jiang Guo", "Shuaichen Chang", "Mingwen Dong", "Dongkyu Lee", "Xing Niu", "Jiarong Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c37dae372655e8ef556163744e47ed42db3533175ca7d1db7560033774ed2a0", "submitted_at": "2025-10-13T19:52:04Z", "updated_at": "2026-03-09T15:25:16Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.11892v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.11892", "identifiers": {"arxiv": "2510.11892"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.11892"}], "name": "R-WoM: Retrieval-augmented World Model For Computer-use Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "R-WoM evaluates how well world models predict nearby GUI states, align complete procedures, and recognize milestone transitions. It retrieves external tutorials to guide simulated future interactions, with experiments on subsets of OSWorld and WebArena.", "topics": ["world-models", "memory", "planning", "evaluation"], "updated_at": "2026-03-09", "url": "https://arxiv.org/abs/2510.11892"}
{"aliases": [], "authors": ["Xiaoxue Ren", "Penghao Jiang", "Kaixin Li", "Zhiyong Huang", "Xiaoning Du", "Jiaojiao Jiang", "Zhenchang Xing", "Jiamou Sun", "Terry Yue Zhuo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e92db1e568c3c512aa00e2363bbdc9869cf834d638c331b45d4753d29a282004", "submitted_at": "2025-10-14T06:52:15Z", "updated_at": "2025-10-14T06:52:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.12200v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.12200", "identifiers": {"arxiv": "2510.12200"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.12200"}], "name": "HackWorld: Evaluating Computer-Use Agents on Exploiting Web Application Vulnerabilities", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces HackWorld for evaluating visual computer-use agents in a capture-the-flag collection of web applications. The benchmark examines interface navigation and multistep planning during vulnerability discovery and exploitation tasks.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2025-10-14", "url": "https://arxiv.org/abs/2510.12200"}
{"aliases": [], "authors": ["Qing Jiang", "Junan Huo", "Xingyu Chen", "Yuda Xiong", "Zhaoyang Zeng", "Yihao Chen", "Tianhe Ren", "Junzhi Yu", "Lei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc05395e60de9bb123ff0418f886b98244bd9273c10d760cc2d40077c6d87868", "submitted_at": "2025-10-14T17:59:54Z", "updated_at": "2025-10-14T17:59:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.12798v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.12798", "identifiers": {"arxiv": "2510.12798"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.12798"}], "name": "Detect Anything via Next Point Prediction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-14", "url": "https://arxiv.org/abs/2510.12798"}
{"aliases": [], "authors": ["Avihay Cohen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bedde658a28c57551c27563fa22cb037839dddd4e5eabe1c66afa27523d0d406", "submitted_at": "2025-10-15T13:39:13Z", "updated_at": "2025-10-15T13:39:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.13543v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.13543", "identifiers": {"arxiv": "2510.13543"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.13543"}], "name": "In-Browser LLM-Guided Fuzzing for Real-Time Prompt Injection Testing in Agentic AI Browsers", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The short abstract supplies the framework and threat setting but no quantitative evaluation details.", "published_at": "2025-10-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes an in-browser fuzzing framework that uses language-model guidance to test indirect prompt-injection vulnerabilities in agentic browsers. The work focuses on unwanted actions triggered by webpage instructions under the browser agent's user privileges.", "topics": ["safety-security", "automation"], "updated_at": "2025-10-15", "url": "https://arxiv.org/abs/2510.13543"}
{"aliases": [], "authors": ["Shrey Pandit", "Xuan-Phi Nguyen", "Yifei Ming", "Austin Xu", "Jiayu Wang", "Caiming Xiong", "Shafiq Joty"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bd91d4bd8eb47c7350ceacd60e2b25e353b31be12fd852d3d7d02d099f62ac1", "submitted_at": "2025-10-15T06:34:46Z", "updated_at": "2025-10-15T06:34:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.13913v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.13913", "identifiers": {"arxiv": "2510.13913"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.13913"}], "name": "Synthesizing Agentic Data for Web Agents with Progressive Difficulty Enhancement Mechanisms", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-15", "url": "https://arxiv.org/abs/2510.13913"}
{"aliases": [], "authors": ["Haiquan Zhao", "Chenhan Yuan", "Fei Huang", "Xiaomeng Hu", "Yichang Zhang", "An Yang", "Bowen Yu", "Dayiheng Liu", "Jingren Zhou", "Junyang Lin", "Baosong Yang", "Chen Cheng", "Jialong Tang", "Jiandong Jiang", "Jianwei Zhang", "Jijie Xu", "Ming Yan", "Minmin Sun", "Pei Zhang", "Pengjun Xie", "Qiaoyu Tang", "Qin Zhu", "Rong Zhang", "Shibin Wu", "Shuo Zhang", "Tao He", "Tianyi Tang", "Tingyu Xia", "Wei Liao", "Weizhou Shen", "Wenbiao Yin", "Wenmeng Zhou", "Wenyuan Yu", "Xiaobin Wang", "Xiaodong Deng", "Xiaodong Xu", "Xinyu Zhang", "Yang Liu", "Yeqiu Li", "Yi Zhang", "Yong Jiang", "Yu Wan", "Yuxin Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43a973ff47a0b656b23e227d4aa70fa4328b00a4d9f74fbdc682307924fe4bc3", "submitted_at": "2025-10-16T04:00:18Z", "updated_at": "2025-10-16T04:00:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.14276v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2510.14276", "identifiers": {"arxiv": "2510.14276"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.14276"}], "name": "Qwen3Guard Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-16", "url": "https://arxiv.org/abs/2510.14276"}
{"aliases": [], "authors": ["Zhe Wu", "Hongjin Lu", "Junliang Xing", "Changhao Zhang", "Yuxuan Li", "Yin Zhu", "Yuhao Yang", "Yuheng Jing", "Kai Li", "Kun Shao", "Jianye Hao", "Jun Wang", "Yuanchun Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95fd3e1286ed49d569b05644d77f822ea9747cfb4769cb7cb93f39c9271cf57e", "submitted_at": "2025-10-16T07:38:21Z", "updated_at": "2026-02-06T02:04:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.14388v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:56Z", "role": "primary", "sha256": "fecd42f8ca75572410c354c892734cb8eefbd3fa622064f6bc6db0f535cf5607", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2510.14388v2"}], "id": "arxiv:2510.14388", "identifiers": {"arxiv": "2510.14388"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.14388"}], "name": "Hi-Agent: Hierarchical Vision-Language Agents for Mobile Device Control", "notes": "Reviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced. This Hi-Agent has a distinct author team and method from the earlier HiAgent working-memory paper; the shared project-name prefix is not treated as an identity match.", "published_at": "2025-10-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Hi-Agent, jointly training a high-level reasoning model and a low-level mobile action model. Execution feedback supplies a foresight advantage for optimizing subgoal decisions, with evaluations on Android control benchmarks.", "topics": ["planning", "reinforcement-learning", "mobile-navigation"], "updated_at": "2026-02-06", "url": "https://arxiv.org/abs/2510.14388"}
{"aliases": [], "authors": ["Rui Wang", "Ce Zhang", "Jun-Yu Ma", "Jianshu Zhang", "Hongru Wang", "Yi Chen", "Boyang Xue", "Tianqing Fang", "Zhisong Zhang", "Hongming Zhang", "Haitao Mi", "Dong Yu", "Kam-Fai Wong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5218a35d29f915c603be550adcb4d8f46d5ebd7176948945fa272187dc3108a3", "submitted_at": "2025-10-16T08:37:42Z", "updated_at": "2026-04-29T01:14:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.14438v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2510.14438", "identifiers": {"arxiv": "2510.14438"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.14438"}], "name": "WebAggregator: Enhancing Compositional Reasoning Capabilities of Deep Research Agent Foundation Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-29", "url": "https://arxiv.org/abs/2510.14438"}
{"aliases": [], "authors": ["Juheon Choi", "Juyong Lee", "Jian Kim", "Chanyoung Kim", "Taywon Min", "W. Bradley Knox", "Min Kyung Lee", "Kimin Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac902743c35419bd79603e55bc8a0cf47028e1974cd07ff43c3c0ada6f3a661a", "submitted_at": "2025-10-16T09:57:35Z", "updated_at": "2026-03-02T09:08:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.14513v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.14513", "identifiers": {"arxiv": "2510.14513"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.14513"}], "name": "State Your Intention to Steer Your Attention: An AI Assistant for Intentional Digital Living", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-02", "url": "https://arxiv.org/abs/2510.14513"}
{"aliases": [], "authors": ["Yuanyi Song", "Heyuan Huang", "Qiqiang Lin", "Yin Zhao", "Xiangmou Qu", "Jun Wang", "Xingyu Lou", "Weiwen Liu", "Zhuosheng Zhang", "Jun Wang", "Yong Yu", "Weinan Zhang", "Zhaoxiang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc45dc62068c971cd831fc3db64bd3d8995d4e5fd60e9c83953476d4fb24f7f3", "submitted_at": "2025-10-16T12:30:05Z", "updated_at": "2025-10-16T12:30:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.14621v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2510.14621", "identifiers": {"arxiv": "2510.14621"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.14621"}], "name": "ColorBench: Benchmarking Mobile Agents with Graph-Structured Framework for Complex Long-Horizon Tasks", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "ColorBench represents mobile tasks as graphs of states collected from real devices, providing a quasi-dynamic environment with multiple valid paths and error paths. It evaluates single-app and cross-app tasks using completion, subtask progress, and atomic diagnostics.", "topics": ["evaluation", "mobile-navigation", "long-horizon", "state-tracking"], "updated_at": "2025-10-16", "url": "https://arxiv.org/abs/2510.14621"}
{"aliases": [], "authors": ["Yiming Wang", "Da Yin", "Yuedong Cui", "Ruichen Zheng", "Zhiqian Li", "Zongyu Lin", "Di Wu", "Xueqing Wu", "Chenchen Ye", "Yu Zhou", "Kai-Wei Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1678260783771fea6cb5f7cdda9b260d341e5fff4f21432092d6cb7d771ad607", "submitted_at": "2025-10-16T17:59:38Z", "updated_at": "2025-10-16T17:59:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.14969v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.14969", "identifiers": {"arxiv": "2510.14969"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.14969"}], "name": "LLMs as Scalable, General-Purpose Simulators For Evolving Digital Agent Training", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-16", "url": "https://arxiv.org/abs/2510.14969"}
{"aliases": [], "authors": ["Gucongcong Fan", "Chaoyue Niu", "Chengfei Lyu", "Fan Wu", "Guihai Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8584797a1c12158b20da04c30e3a0f577c63f02f9e906a051579959de2cae480", "submitted_at": "2025-10-17T09:11:05Z", "updated_at": "2025-10-17T09:11:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.15455v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.15455", "identifiers": {"arxiv": "2510.15455"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.15455"}], "name": "CORE: Reducing UI Exposure in Mobile Agents via Collaboration Between Cloud and Local LLMs", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. XML screen hierarchies do not by themselves establish Android deployment.", "published_at": "2025-10-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CORE, which divides mobile UI hierarchies into semantic blocks and shares planning between local and cloud language models. Local ranking limits the screen information sent for cloud-side element selection, with a multi-round mechanism for correcting inadequate context.", "topics": ["safety-security", "mobile-navigation", "efficiency"], "updated_at": "2025-10-17", "url": "https://arxiv.org/abs/2510.15455"}
{"aliases": [], "authors": ["Simon Yu", "Gang Li", "Weiyan Shi", "Peng Qi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "864f797fd3d3ea317ca29d225d67db3ea635f840afdb30639b4bd0368bc5ac17", "submitted_at": "2025-10-17T17:56:00Z", "updated_at": "2026-03-02T03:34:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.15863v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.15863", "identifiers": {"arxiv": "2510.15863"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.15863"}], "name": "PolySkill: Learning Generalizable Skills Through Polymorphic Abstraction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-02", "url": "https://arxiv.org/abs/2510.15863"}
{"aliases": [], "authors": ["Sofiya Garkot", "Maksym Shamrai", "Ivan Synytsia", "Mariya Hirna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e4377fd6b510455afa1eb04c64e8e0b551e89d3b54ce31f241f283233985e1c", "submitted_at": "2025-10-16T19:03:45Z", "updated_at": "2026-03-24T14:39:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.16051v2"}}, "environments": ["macos", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2510.16051", "identifiers": {"arxiv": "2510.16051"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.16051"}], "name": "GUIrilla: A Scalable Framework for Automated Desktop UI Exploration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUIrilla systematically explores macOS interfaces to collect interaction traces and accessibility metadata for downstream GUI research. It organizes these observations into hierarchical MacApp Trees and describes an accessibility-driven collection library.", "topics": ["data-generation", "screen-understanding", "recording"], "updated_at": "2026-03-24", "url": "https://arxiv.org/abs/2510.16051"}
{"aliases": [], "authors": ["Yueqian Lin", "Zhengmian Hu", "Jayakumar Subramanian", "Qinsi Wang", "Nikos Vlassis", "Hai \"Helen\" Li", "Yiran Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9cb8554bc968e3a2a30ff3bf63dadf76fde404a1136f063ad834bec1911100d9", "submitted_at": "2025-10-17T19:00:08Z", "updated_at": "2025-10-17T19:00:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.16156v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.16156", "identifiers": {"arxiv": "2510.16156"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.16156"}], "name": "AsyncVoice Agent: Real-Time Explanation for LLM Planning and Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-17", "url": "https://arxiv.org/abs/2510.16156"}
{"aliases": [], "authors": ["Yuxuan Lu", "Ziyi Wang", "Jing Huang", "Hui Liu", "Jiri Gesi", "Yan Han", "Shihan Fu", "Tianqi Zheng", "Xianfeng Tang", "Chen Luo", "Yisi Sang", "Jin Lai", "Dakuo Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9a1f8ebdf70ff2fa7b0c807160cd583c0f017a536457ff5915e589fca5fa751", "submitted_at": "2025-10-17T22:54:33Z", "updated_at": "2026-05-17T17:26:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.16252v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.16252", "identifiers": {"arxiv": "2510.16252"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.16252"}], "name": "WEBSERV: A Full-Stack and RL-Ready Web Environment for Training Web Agents at Scale", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebServ combines copy-on-write container environments with a compact DOM-derived browser interface for training web agents. Network-aware waiting supports action execution in single-page applications, with evaluations and reinforcement-learning experiments on WebArena-Lite.", "topics": ["sandbox", "reinforcement-learning", "efficiency", "web-navigation"], "updated_at": "2026-05-17", "url": "https://arxiv.org/abs/2510.16252"}
{"aliases": [], "authors": ["Qiusi Zhan", "Angeline Budiman-Chan", "Abdelrahman Zayed", "Xingzhi Guo", "Daniel Kang", "Joo-Kyung Kim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "172bccc0cfb385618e6357e7b84e32c0a7dc5a22e1e5f4b5a518601bbab8f18e", "submitted_at": "2025-10-19T21:47:19Z", "updated_at": "2026-03-20T19:17:30Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2510.17017v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.17017", "identifiers": {"arxiv": "2510.17017"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.17017"}], "name": "SafeSearch: Do Not Trade Safety for Utility in LLM Search Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-20", "url": "https://arxiv.org/abs/2510.17017"}
{"aliases": [], "authors": ["Yuhao Yang", "Zhen Yang", "Zi-Yi Dou", "Anh Nguyen", "Keen You", "Omar Attia", "Andrew Szot", "Michael Feng", "Ram Ramrakhya", "Alexander Toshev", "Chao Huang", "Yinfei Yang", "Zhe Gan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33569f6ccca9c7e2675796191124424e76636828179c300d35da01ad9ccef1ce", "submitted_at": "2025-10-20T17:48:26Z", "updated_at": "2026-05-26T11:29:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.17790v3"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.17790", "identifiers": {"arxiv": "2510.17790"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.17790"}], "name": "UltraCUA: A Foundation Model for Computer Use Agents with Hybrid Action", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "UltraCUA trains a shared policy for primitive GUI actions and high-level tool calls using synthesized tasks and hybrid trajectories. Supervised fine-tuning followed by online reinforcement learning is evaluated on OSWorld and WindowsAgentArena.", "topics": ["tool-integration", "reinforcement-learning", "data-generation"], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2510.17790"}
{"aliases": [], "authors": ["Jiale Cheng", "Yusen Liu", "Xinyu Zhang", "Yulin Fei", "Wenyi Hong", "Ruiliang Lyu", "Weihan Wang", "Zhe Su", "Xiaotao Gu", "Xiao Liu", "Yushi Bai", "Jie Tang", "Hongning Wang", "Minlie Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a4f41eefe43dac9c775789a23ed3a05425655400095df27458d7925961ebd70", "submitted_at": "2025-10-20T17:58:56Z", "updated_at": "2025-10-21T17:12:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.17800v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.17800", "identifiers": {"arxiv": "2510.17800"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.17800"}], "name": "Glyph: Scaling Context Windows via Visual-Text Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-21", "url": "https://arxiv.org/abs/2510.17800"}
{"aliases": [], "authors": ["Devin Ersoy", "Brandon Lee", "Ananth Shreekumar", "Arjun Arunasalam", "Muhammad Ibrahim", "Antonio Bianchi", "Z. Berkay Celik"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db3c25ba55221ca3d7c45ce4243a4a37b56a698909c0ea3eaaabdcd37388a1ec", "submitted_at": "2025-10-20T21:26:26Z", "updated_at": "2025-10-20T21:26:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.18113v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.18113", "identifiers": {"arxiv": "2510.18113"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.18113"}], "name": "Investigating the Impact of Dark Patterns on LLM-Based Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents LiteAgent for recording web-agent interactions and TrickyArena for enabling or disabling dark patterns in controlled websites. Experiments compare how individual patterns, combined patterns and visual or HTML changes affect agent decisions.", "topics": ["safety-security", "ui-design", "evaluation"], "updated_at": "2025-10-20", "url": "https://arxiv.org/abs/2510.18113"}
{"aliases": [], "authors": ["Zheng Zhang", "Jiarui He", "Yuchen Cai", "Deheng Ye", "Peilin Zhao", "Ruili Feng", "Hao Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "30123a470c2caa4b755ea696c9dd50d501e8d627e5b2cf1c80801807cd78a1f8", "submitted_at": "2025-10-21T05:49:37Z", "updated_at": "2026-04-01T16:25:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.18314v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.18314", "identifiers": {"arxiv": "2510.18314"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.18314"}], "name": "Genesis: Evolving Attack Strategies for LLM Web Agent Red-Teaming", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-01", "url": "https://arxiv.org/abs/2510.18314"}
{"aliases": [], "authors": ["Ho Fai Leung", "Xiaoyan Xi", "Fei Zuo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33b21a63d51148f2b5c30478c302e3df9b92e03e2e106a1c68224f2c835a8608", "submitted_at": "2025-10-21T10:11:33Z", "updated_at": "2025-10-21T10:11:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.18488v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.18488", "identifiers": {"arxiv": "2510.18488"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.18488"}], "name": "AndroidControl-Curated: Revealing the True Potential of GUI Agents through Benchmark Purification", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "AndroidControl-Curated audits ambiguities and factual errors in AndroidControl and produces a refined evaluation dataset. The paper also studies post-training a small Magma-R1 model on a curated subset to examine the effect of data quality on GUI action assessment.", "topics": ["evaluation", "data-generation", "grounding"], "updated_at": "2025-10-21", "url": "https://arxiv.org/abs/2510.18488"}
{"aliases": [], "authors": ["Haojia Lin", "Xiaoyu Tan", "Yulei Qin", "Zihan Xu", "Yuchen Shi", "Zongyi Li", "Gang Li", "Shaofei Cai", "Siqi Cai", "Chaoyou Fu", "Ke Li", "Xing Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73e11cfa52e6ad930f44147ffac6f54f93a95a7f6992371686acff7f5cf4ed06", "submitted_at": "2025-10-21T12:53:40Z", "updated_at": "2025-10-21T12:53:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.18596v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.18596", "identifiers": {"arxiv": "2510.18596"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.18596"}], "name": "CUARewardBench: A Benchmark for Evaluating Reward Models on Computer-using Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-21", "url": "https://arxiv.org/abs/2510.18596"}
{"aliases": [], "authors": ["Yaoyao Qian", "Yuanli Wang", "Jinda Zhang", "Yun Zong", "Meixu Chen", "Hanhan Zhou", "Jindan Huang", "Yifan Zeng", "Xinyu Hu", "Chan Hee Song", "Danqing Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d404ff922ef38e57481c01a5e3108bd38457d88f83d7adf6d3c3c7b5ddf124e2", "submitted_at": "2025-10-22T03:29:25Z", "updated_at": "2025-10-22T03:29:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.19205v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.19205", "identifiers": {"arxiv": "2510.19205"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.19205"}], "name": "WebGraphEval: Multi-Turn Trajectory Evaluation for Web Agents using Graph Representation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebGraphEval, which merges multiple agents' trajectories into a weighted graph of canonical actions. Structural analyses examine recurring behavior, inefficient paths and important decision points beyond a final success score.", "topics": ["evaluation", "observability", "interaction-history"], "updated_at": "2025-10-22", "url": "https://arxiv.org/abs/2510.19205"}
{"aliases": [], "authors": ["Yizhu Wang", "Sizhe Chen", "Raghad Alkhudair", "Basel Alomair", "David Wagner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a44ad3762b1b547e22d24b45197ed887b1c57df1227cf89912700b0f0cc682d", "submitted_at": "2025-10-22T03:30:49Z", "updated_at": "2026-02-04T07:26:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.19207v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.19207", "identifiers": {"arxiv": "2510.19207"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.19207"}], "name": "Defending Against Prompt Injection with DataFilter", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-04", "url": "https://arxiv.org/abs/2510.19207"}
{"aliases": [], "authors": ["Ziwei Wang", "Jiayuan Su", "Mengyu Zhou", "Huaxing Zeng", "Mengni Jia", "Xiao Lv", "Haoyu Dong", "Xiaojun Ma", "Shi Han", "Dongmei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e443bf41effee1ff02b35f973d5a412ea86298757f35c9c0068fa6cc172b21f1", "submitted_at": "2025-10-22T05:09:44Z", "updated_at": "2025-10-22T05:09:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.19247v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.19247", "identifiers": {"arxiv": "2510.19247"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.19247"}], "name": "SheetBrain: A Neuro-Symbolic Agent for Accurate Reasoning over Complex and Large Spreadsheets", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-22", "url": "https://arxiv.org/abs/2510.19247"}
{"aliases": [], "authors": ["Ning Li", "Qiqiang Lin", "Zheng Wu", "Xiaoyun Mo", "Weiming Zhang", "Yin Zhao", "Xiangmou Qu", "Jiamu Zhou", "Jun Wang", "Congmin Zheng", "Yuanyi Song", "Hongjiang Chen", "Heyuan Huang", "Jihong Wang", "Jiaxin Yin", "Jingwei Yu", "Junwei Liao", "Qiuying Peng", "Xingyu Lou", "Jun Wang", "Weiwen Liu", "Zhuosheng Zhang", "Weinan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2ccedef5abd1b7219b1f0e05e4532519aed811de02eb763cfc4452a131444da", "submitted_at": "2025-10-22T09:02:48Z", "updated_at": "2025-10-24T07:32:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.19386v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.19386", "identifiers": {"arxiv": "2510.19386"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.19386"}], "name": "ColorAgent: Building A Robust, Personalized, and Interactive OS Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-24", "url": "https://arxiv.org/abs/2510.19386"}
{"aliases": [], "authors": ["Dunjie Lu", "Yiheng Xu", "Junli Wang", "Haoyuan Wu", "Xinyuan Wang", "Zekun Wang", "Junlin Yang", "Hongjin Su", "Jixuan Chen", "Junda Chen", "Yuchen Mao", "Jingren Zhou", "Junyang Lin", "Binyuan Hui", "Tao Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebb064e1a9c3e2255677d9b4cd9ce33d2a97dd26a379c321721d576e7c5d6399", "submitted_at": "2025-10-22T11:25:48Z", "updated_at": "2025-10-22T11:25:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.19488v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.19488", "identifiers": {"arxiv": "2510.19488"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.19488"}], "name": "VideoAgentTrek: Computer Use Pretraining from Unlabeled Videos", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Extracts structured GUI actions from screen-recorded tutorials using inverse dynamics, then uses the inferred trajectories for computer-use training.", "topics": ["data-generation", "video"], "updated_at": "2025-10-22", "url": "https://arxiv.org/abs/2510.19488"}
{"aliases": [], "authors": ["Amila Indika", "Igor Molybog"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6dcf5b31dbb6dd5040b9d8d4f7bd7dfbb5d442cc7c47b1a0fb10adf5eae1f988", "submitted_at": "2025-10-22T01:36:13Z", "updated_at": "2025-10-22T01:36:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.19864v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.19864", "identifiers": {"arxiv": "2510.19864"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.19864"}], "name": "SODBench: A Large Language Model Approach to Documenting Spreadsheet Operations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-22", "url": "https://arxiv.org/abs/2510.19864"}
{"aliases": [], "authors": ["Mathieu Andreux", "Märt Bakler", "Yanael Barbier", "Hamza Benchekroun", "Emilien Biré", "Antoine Bonnet", "Riaz Bordie", "Nathan Bout", "Matthias Brunel", "Aleix Cambray", "Pierre-Louis Cedoz", "Antoine Chassang", "Gautier Cloix", "Ethan Connelly", "Alexandra Constantinou", "Ramzi De Coster", "Hubert de la Jonquiere", "Aurélien Delfosse", "Maxime Delpit", "Alexis Deprez", "Augustin Derupti", "Mathieu Diaz", "Shannon D'Souza", "Julie Dujardin", "Abai Edmund", "Michael Eickenberg", "Armand Fatalot", "Wissem Felissi", "Isaac Herring", "Xavier Koegler", "Erwan Le Jumeau de Kergaradec", "Aurélien Lac", "Maxime Langevin", "Corentin Lauverjat", "Antonio Loison", "Avshalom Manevich", "Axel Moyal", "Axel Nguyen Kerbel", "Marinela Parovic", "Julien Revelle", "Guillaume Richard", "Mats Richter", "Ronan Riochet", "María Santos", "Romain Savidan", "Laurent Sifre", "Maxime Theillard", "Marc Thibault", "Ivan Valentini", "Tony Wu", "Laura Yie", "Kai Yuan", "Jevgenij Zubovskij"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de25494af1bc2ceaa760a21da51803aaf89d9d2d1961887cfd0023fb87f28f63", "submitted_at": "2025-10-22T18:21:52Z", "updated_at": "2025-10-24T11:52:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.19949v2"}}, "environments": ["web", "desktop", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.19949", "identifiers": {"arxiv": "2510.19949"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.19949"}], "name": "Surfer 2: The Next Generation of Cross-Platform Computer Use Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Surfer2 combines hierarchical context management with separate planning and execution, self-verification, and recovery for visual computer use. It evaluates the framework across web, desktop, and Android tasks without task-specific fine-tuning.", "topics": ["planning", "memory", "long-horizon", "robustness"], "updated_at": "2025-10-24", "url": "https://arxiv.org/abs/2510.19949"}
{"aliases": [], "authors": ["Liangyu Chen", "Hanzhang Zhou", "Chenglin Cai", "Jianan Zhang", "Panrong Tong", "Quyu Kong", "Xu Zhang", "Chen Liu", "Yuqi Liu", "Wenxuan Wang", "Yue Wang", "Qin Jin", "Steven Hoi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3378dd9187f8c8df983d97b10b2ca5b64b2e2332760a734a09db8ce34a398a0b", "submitted_at": "2025-10-23T07:18:32Z", "updated_at": "2025-10-23T07:18:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.20286v1"}}, "environments": ["cross-platform", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.20286", "identifiers": {"arxiv": "2510.20286"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.20286"}], "name": "UI-Ins: Enhancing GUI Grounding with Multi-Perspective Instruction-as-Reasoning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Treats alternative instruction formulations as reasoning paths and trains GUI-grounding models to select and combine useful perspectives.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-10-23", "url": "https://arxiv.org/abs/2510.20286"}
{"aliases": [], "authors": ["Chiyu Chen", "Xinhao Song", "Yunkai Chai", "Yang Yao", "Haodong Zhao", "Lijun Li", "Jie Li", "Yan Teng", "Gongshen Liu", "Yingchun Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8582fa41bb5f4d5c464c3bb74285a4ef17545cde34e48e4a1b36cd35937337ae", "submitted_at": "2025-10-23T08:33:24Z", "updated_at": "2026-03-05T06:51:24Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.20333v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.20333", "identifiers": {"arxiv": "2510.20333"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.20333"}], "name": "GhostEI-Bench: Do Mobile Agents Resilience to Environmental Injection in Dynamic On-Device Environments?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-05", "url": "https://arxiv.org/abs/2510.20333"}
{"aliases": [], "authors": ["Jonathan Bragg", "Mike D'Arcy", "Nishant Balepur", "Dan Bareket", "Bhavana Dalvi", "Sergey Feldman", "Dany Haddad", "Jena D. Hwang", "Peter Jansen", "Varsha Kishore", "Bodhisattwa Prasad Majumder", "Aakanksha Naik", "Sigal Rahamimov", "Kyle Richardson", "Amanpreet Singh", "Harshit Surana", "Aryeh Tiktinsky", "Rosni Vasu", "Guy Wiener", "Chloe Anastasiades", "Stefan Candra", "Jason Dunkelberger", "Dan Emery", "Rob Evans", "Malachi Hamada", "Regan Huff", "Rodney Kinney", "Matt Latzke", "Jaron Lochner", "Ruben Lozano-Aguilera", "Cecile Nguyen", "Smita Rao", "Amber Tanaka", "Brooke Vlahos", "Peter Clark", "Doug Downey", "Yoav Goldberg", "Ashish Sabharwal", "Daniel S. Weld"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2767d4d11e6f3ab54232d093fa84c96d35ff01f0009f2be7b06f51d114fc249f", "submitted_at": "2025-10-24T17:10:26Z", "updated_at": "2026-04-21T23:10:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.21652v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.21652", "identifiers": {"arxiv": "2510.21652"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.21652"}], "name": "AstaBench: Rigorous Benchmarking of AI Agents with a Scientific Research Suite", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-21", "url": "https://arxiv.org/abs/2510.21652"}
{"aliases": [], "authors": ["Yangqin Jiang", "Chao Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "09bd73dff01641593f31f7d40effd60df42e3d9460ffa010f76c00a32be3593b", "submitted_at": "2025-10-24T20:23:12Z", "updated_at": "2026-10-05T01:17:12Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.22009v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2510.22009", "identifiers": {"arxiv": "2510.22009"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.22009"}], "name": "OpenPhone: Mobile Agentic Foundation Models", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "OpenPhone trains a compact mobile policy with supervised learning followed by GRPO on synthetic GUI data, incorporating reasoning and memory management. Its device-cloud design handles routine work locally and delegates complex subtasks, with online evaluation on AndroidLab.", "topics": ["reinforcement-learning", "memory", "efficiency", "mobile-navigation"], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2510.22009"}
{"aliases": [], "authors": ["Vijay Veerabadran", "Fanyi Xiao", "Nitin Kamra", "Pedro Matias", "Joy Chen", "Caley Drooff", "Brett D Roads", "Riley Williams", "Ethan Henderson", "Xuanyi Zhao", "Kevin Carlberg", "Joseph Tighe", "Karl Ridgeway"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a35e8ecfa09570c24c16c85debf24b7f9ca4989e6694622e7de5086802aef5a", "submitted_at": "2025-10-25T21:54:01Z", "updated_at": "2025-10-25T21:54:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.22443v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.22443", "identifiers": {"arxiv": "2510.22443"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.22443"}], "name": "Benchmarking Egocentric Multimodal Goal Inference for Assistive Wearable Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-25", "url": "https://arxiv.org/abs/2510.22443"}
{"aliases": [], "authors": ["Jiali Cheng", "Anjishnu Kumar", "Roshan Lal", "Rishi Rajasekaran", "Hani Ramezani", "Omar Zia Khan", "Oleg Rokhlenko", "Sunny Chiu-Webster", "Gang Hua", "Hadi Amiri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff353afdf50c851cbc3a5c29a068b7403712bc87e0e9b5ec1c75d8fae9e5745d", "submitted_at": "2025-10-26T16:03:39Z", "updated_at": "2025-12-19T23:16:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.22732v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "id": "arxiv:2510.22732", "identifiers": {"arxiv": "2510.22732"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.22732"}], "name": "WebATLAS: An LLM Agent with Experience-Driven Memory and Action Simulation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-19", "url": "https://arxiv.org/abs/2510.22732"}
{"aliases": [], "authors": ["Mouxiao Huang", "Borui Jiang", "Dehua Zheng", "Hailin Hu", "Kai Han", "Xinghao Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73a5fd857411428c5b9f8e53c7c6200452a1e06ad850f31dc99e1e14f798462e", "submitted_at": "2025-10-27T02:40:02Z", "updated_at": "2026-02-05T08:25:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.22936v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.22936", "identifiers": {"arxiv": "2510.22936"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.22936"}], "name": "PPE: Positional Preservation Embedding for Token Compression in Multimodal Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-05", "url": "https://arxiv.org/abs/2510.22936"}
{"aliases": [], "authors": ["Litu Ou", "Kuan Li", "Huifeng Yin", "Liwen Zhang", "Zhongwang Zhang", "Xixi Wu", "Rui Ye", "Zile Qiao", "Pengjun Xie", "Jingren Zhou", "Yong Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "042a4e25e1440d3716a943cd1fc6487279b7c35ff9b44481738bc8f3161f1ba7", "submitted_at": "2025-10-27T15:58:51Z", "updated_at": "2025-10-28T16:23:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.23458v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.23458", "identifiers": {"arxiv": "2510.23458"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.23458"}], "name": "BrowseConf: Confidence-Guided Test-Time Scaling for Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-28", "url": "https://arxiv.org/abs/2510.23458"}
{"aliases": [], "authors": ["Scott Jones", "Prakash Murali"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f64e5bdf959da65ed0d2f90a8bbb2e600f7cddd9122ee468137108c00722bb0d", "submitted_at": "2025-10-27T16:58:00Z", "updated_at": "2026-04-30T21:36:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.23519v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.23519", "identifiers": {"arxiv": "2510.23519"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.23519"}], "name": "Architecting Scalable Trapped Ion Quantum Computers using Surface Codes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-30", "url": "https://arxiv.org/abs/2510.23519"}
{"aliases": [], "authors": ["Alessandra Terranova", "Björn Ross", "Alexandra Birch"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d79d9928728017f7f21b50ca823fd9df678185426b9cd36eb03f3a0d3bff7324", "submitted_at": "2025-10-27T18:03:50Z", "updated_at": "2025-12-06T18:19:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.23730v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.23730", "identifiers": {"arxiv": "2510.23730"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.23730"}], "name": "Evaluating Long-Term Memory for Long-Context Question Answering", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-06", "url": "https://arxiv.org/abs/2510.23730"}
{"aliases": [], "authors": ["Anshuman Chhabra", "Shrestha Datta", "Shahriar Kabir Nahin", "Prasant Mohapatra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d908843d49f2718bf5cd405fb37003496971d7e9a0cb3f2b93c94c5d1e40a46c", "submitted_at": "2025-10-27T21:48:11Z", "updated_at": "2026-04-03T16:27:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.23883v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.23883", "identifiers": {"arxiv": "2510.23883"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.23883"}], "name": "Agentic AI Security: Threats, Defenses, Evaluation, and Open Challenges", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-03", "url": "https://arxiv.org/abs/2510.23883"}
{"aliases": [], "authors": ["Weihua Cheng", "Junming Liu", "Yifei Sun", "Botian Shi", "Yirong Chen", "Ding Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d80312fbd8f57cdc53cc10e03f2c68e2152d06945e5089967dceda50c02b9d4b", "submitted_at": "2025-10-28T08:19:58Z", "updated_at": "2026-04-14T04:34:20Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.24168v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.24168", "identifiers": {"arxiv": "2510.24168"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.24168"}], "name": "MGA: Memory-Driven GUI Agent for Observation-Centric Interaction", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-10-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MGA, which separates task-agnostic screen-state reading from decision making and connects steps through structured state memory. Validated state deltas replace full trajectory replay, evaluated on OSWorld and application tasks.", "topics": ["memory", "state-tracking", "long-horizon"], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2510.24168"}
{"aliases": [], "authors": ["Qiushi Sun", "Mukai Li", "Zhoumianze Liu", "Zhihui Xie", "Fangzhi Xu", "Zhangyue Yin", "Kanzhi Cheng", "Zehao Li", "Zichen Ding", "Qi Liu", "Zhiyong Wu", "Zhuosheng Zhang", "Ben Kao", "Lingpeng Kong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e3fde732c9e760437f2c60d585f7ba0d2ed13bbe9e051838844a21174159099", "submitted_at": "2025-10-28T13:22:39Z", "updated_at": "2026-07-26T01:16:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.24411v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.24411", "identifiers": {"arxiv": "2510.24411"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.24411"}], "name": "OS-Sentinel: Towards Safety-Enhanced Mobile GUI Agents via Hybrid Validation in Realistic Workflows", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-10-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MobileRisk-Live and OS-Sentinel, combining a formal verifier for system-level violations with a multimodal judge for contextual risks in mobile GUI workflows.", "topics": ["safety-security", "evaluation", "mobile-navigation"], "updated_at": "2026-07-26", "url": "https://arxiv.org/abs/2510.24411"}
{"aliases": [], "authors": ["Hongrui Jia", "Jitong Liao", "Xi Zhang", "Haiyang Xu", "Tianbao Xie", "Chaoya Jiang", "Ming Yan", "Si Liu", "Wei Ye", "Fei Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d115356f723746ce95c3e40fed933786a0747b3b560c8f73a6b7af05ca22d24f", "submitted_at": "2025-10-28T15:56:36Z", "updated_at": "2025-11-11T04:10:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.24563v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.24563", "identifiers": {"arxiv": "2510.24563"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.24563"}], "name": "OSWorld-MCP: Benchmarking MCP Tool Invocation In Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "OSWorld-MCP evaluates how agents choose between GUI operations and MCP tool calls in computer tasks. It combines generated and curated tools, with 158 manually validated tools across seven applications, to examine tool use alongside task completion.", "topics": ["evaluation", "tool-integration", "planning"], "updated_at": "2025-11-11", "url": "https://arxiv.org/abs/2510.24563"}
{"aliases": [], "authors": ["Rui Ye", "Zhongwang Zhang", "Kuan Li", "Huifeng Yin", "Zhengwei Tao", "Yida Zhao", "Liangcai Su", "Liwen Zhang", "Zile Qiao", "Xinyu Wang", "Pengjun Xie", "Fei Huang", "Siheng Chen", "Jingren Zhou", "Yong Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a03cc707ceb15847a8fd07197d6159bb308d7024690af3ef14676e6d6638727f", "submitted_at": "2025-10-28T17:51:50Z", "updated_at": "2025-10-28T17:51:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.24699v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.24699", "identifiers": {"arxiv": "2510.24699"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.24699"}], "name": "AgentFold: Long-Horizon Web Agents with Proactive Context Management", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-28", "url": "https://arxiv.org/abs/2510.24699"}
{"aliases": [], "authors": ["Yueqi Song", "Ketan Ramaneti", "Zaid Sheikh", "Ziru Chen", "Boyu Gou", "Tianbao Xie", "Yiheng Xu", "Danyang Zhang", "Apurva Gandhi", "Fan Yang", "Joseph Liu", "Tianyue Ou", "Zhihao Yuan", "Frank Xu", "Shuyan Zhou", "Xingyao Wang", "Xiang Yue", "Tao Yu", "Huan Sun", "Yu Su", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1babedca47327517bc12cd93ea8956180f4e6aa55f7104201019e9c31bb5043d", "submitted_at": "2025-10-28T17:53:13Z", "updated_at": "2026-03-04T04:17:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.24702v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.24702", "identifiers": {"arxiv": "2510.24702"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.24702"}], "name": "Agent Data Protocol: Unifying Datasets for Diverse, Effective Fine-tuning of LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-04", "url": "https://arxiv.org/abs/2510.24702"}
{"aliases": [], "authors": ["Junlong Li", "Wenshuo Zhao", "Jian Zhao", "Weihao Zeng", "Haoze Wu", "Xiaochen Wang", "Rui Ge", "Yuxuan Cao", "Yuzhen Huang", "Wei Liu", "Junteng Liu", "Zhaochen Su", "Yiyang Guo", "Fan Zhou", "Lueyang Zhang", "Juan Michelini", "Xingyao Wang", "Xiang Yue", "Shuyan Zhou", "Graham Neubig", "Junxian He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5cf3a6ee05919f7a939ccf7c52fc9961aadced0761b47b7de255576bfee8a6cb", "submitted_at": "2025-10-29T17:32:49Z", "updated_at": "2026-02-26T09:46:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.25726v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.25726", "identifiers": {"arxiv": "2510.25726"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.25726"}], "name": "The Tool Decathlon: Benchmarking Language Agents for Diverse, Realistic, and Long-Horizon Task Execution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-26", "url": "https://arxiv.org/abs/2510.25726"}
{"aliases": [], "authors": ["David Schmotz", "Sahar Abdelnabi", "Maksym Andriushchenko"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "72afd3c38b32e43bdd115a122d1e55e76631b9792b081e332e0f2708b30a1d96", "submitted_at": "2025-10-30T10:27:11Z", "updated_at": "2025-10-30T10:27:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.26328v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.26328", "identifiers": {"arxiv": "2510.26328"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.26328"}], "name": "Agent Skills Enable a New Class of Realistic and Trivially Simple Prompt Injections", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-30", "url": "https://arxiv.org/abs/2510.26328"}
{"aliases": [], "authors": ["Shao-Hua Hu", "Po-Sung Liu", "Jun-Yi Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a931d8151e40507bed801bb7fcc36d5a2ff3c782d34bb4e7dbea175b70781524", "submitted_at": "2025-10-30T17:58:13Z", "updated_at": "2026-04-16T01:46:35Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.26789v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.26789", "identifiers": {"arxiv": "2510.26789"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.26789"}], "name": "Entanglement-assisted circuit knitting: Distributed quantum computing using limited entanglement resources", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-16", "url": "https://arxiv.org/abs/2510.26789"}
{"aliases": [], "authors": ["Tong Shen", "Jingai Yu", "Dong Zhou", "Dong Li", "Emad Barsoum"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7351368136da6ce50fd188983c7c88bba27571a51c1f25f8c895747e72c969f", "submitted_at": "2025-10-31T03:13:08Z", "updated_at": "2025-10-31T03:13:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.27135v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.27135", "identifiers": {"arxiv": "2510.27135"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.27135"}], "name": "E-MMDiT: Revisiting Multimodal Diffusion Transformer Design for Fast Image Synthesis under Limited Resources", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-10-31", "url": "https://arxiv.org/abs/2510.27135"}
{"aliases": [], "authors": ["Chenghao Du", "Quanfeng Huang", "Tingxuan Tang", "Zihao Wang", "Adwait Nadkarni", "Yue Xiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61b611712fff54af25aae660f2353fa6c3aaa764b0767831ebf2278c120d9dea", "submitted_at": "2025-10-31T03:35:59Z", "updated_at": "2026-07-07T02:18:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.27140v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2510.27140", "identifiers": {"arxiv": "2510.27140"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.27140"}], "name": "Measuring the Security of Mobile LLM Agents under Adversarial Prompts from Untrusted Third-Party Channels", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Measures how untrusted third-party notifications, messages, and app content can hijack mobile LLM agents into executing unauthorized cross-app actions.", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2026-07-07", "url": "https://arxiv.org/abs/2510.27140"}
{"aliases": [], "authors": ["Tao Liu", "Chongyu Wang", "Rongjie Li", "Yingchen Yu", "Xuming He", "Bai Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "283b1a1de2a4e1d4ec8175f0c15af66c10097a7f5ad9fd2793e93b888526ff7e", "submitted_at": "2025-10-31T06:10:57Z", "updated_at": "2025-10-31T06:10:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.27210v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2510.27210", "identifiers": {"arxiv": "2510.27210"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.27210"}], "name": "GUI-Rise: Structured Reasoning and History Summarization for GUI Navigation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-10-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines structured progress-and-decision reasoning with history summarization to improve cross-domain generalization in GUI navigation agents.", "topics": ["planning", "interaction-history", "reinforcement-learning"], "updated_at": "2025-10-31", "url": "https://arxiv.org/abs/2510.27210"}
{"aliases": [], "authors": ["Shaojie Zhang", "Pei Fu", "Ruoceng Zhang", "Jiahui Yang", "Anan Du", "Xiuwen Xi", "Shaokang Wang", "Ying Huang", "Bin Qin", "Zhenbo Luo", "Jian Luan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6869e86d7cd4242ba8dcf942bb1cef3c157dff4f0ef37f7138ff5025d082c7b8", "submitted_at": "2025-10-31T08:07:02Z", "updated_at": "2026-05-27T02:32:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.27266v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2510.27266", "identifiers": {"arxiv": "2510.27266"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.27266"}], "name": "Enhancing Trustworthy GUI Grounding via Self-Critiqued Reinforcement Learning", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-10-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "HyperClick trains GUI grounders to produce a click and an accompanying confidence estimate using rewards for correctness and confidence alignment. Its evaluation examines localization, calibration, and abstention behavior.", "topics": ["grounding", "reinforcement-learning", "robustness", "evaluation"], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2510.27266"}
{"aliases": [], "authors": ["Jianwen Sun", "Yukang Feng", "Yifan Chang", "Chuanhao Li", "Zizhen Li", "Jiaxin Ai", "Fanrui Zhang", "Yu Dai", "Kaipeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c3c781f1c4d425c766e25fc7b90a1d34dc64f2a9f6eda2b5b0bfdde479c6bad", "submitted_at": "2025-10-31T12:00:21Z", "updated_at": "2026-01-30T06:09:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.27410v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2510.27410", "identifiers": {"arxiv": "2510.27410"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2510.27410"}], "name": "Closing the Expression Gap in LLM Instructions via Socratic Questioning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-30", "url": "https://arxiv.org/abs/2510.27410"}
{"aliases": [], "authors": ["Justin Yu", "Yide Shentu", "Di Wu", "Pieter Abbeel", "Ken Goldberg", "Philipp Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3dd8dc9ed3232280c47cc77bc30e2a97ef7d2fccdfce36da5f2ad23c175edf3f", "submitted_at": "2025-10-31T18:00:35Z", "updated_at": "2026-03-10T00:29:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.00153v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.00153", "identifiers": {"arxiv": "2511.00153"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.00153"}], "name": "EgoMI: Learning Active Vision and Whole-Body Manipulation from Egocentric Human Demonstrations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-10-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-10", "url": "https://arxiv.org/abs/2511.00153"}
{"aliases": [], "authors": ["Shijie Zhou", "Viet Dac Lai", "Hao Tan", "Jihyung Kil", "Wanrong Zhu", "Changyou Chen", "Ruiyi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c2224c58c8fbac7a68f3757391f5e6a4f62c32fa42aae86bf9821c0755b402b1", "submitted_at": "2025-11-02T05:34:21Z", "updated_at": "2026-06-30T10:43:39Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2511.00810v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.00810", "identifiers": {"arxiv": "2511.00810"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.00810"}], "name": "GUI-AIMA: Aligning Intrinsic Multimodal Attention with a Context Anchor for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI-AIMA, which supervises instruction-relevant attention over visual patches instead of directly training coordinate text generation. Patch-level grounding can be combined with an optional zoom stage and is evaluated across GUI localization benchmarks.", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2026-06-30", "url": "https://arxiv.org/abs/2511.00810"}
{"aliases": [], "authors": ["Tianfan Peng", "Yuntao Du", "Pengzhou Ji", "Shijie Dong", "Kailin Jiang", "Mingchuan Ma", "Yijun Tian", "Jinhe Bi", "Qian Li", "Wei Du", "Feng Xiao", "Lizhen Cui"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "851940f3ba12ac858b875a535e5979a9651bb328fc638d7c3f862089188a532f", "submitted_at": "2025-11-04T15:17:06Z", "updated_at": "2025-11-15T12:23:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.02650v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.02650", "identifiers": {"arxiv": "2511.02650"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.02650"}], "name": "Can Visual Input Be Compressed? A Visual Token Compression Benchmark for Large Multimodal Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-15", "url": "https://arxiv.org/abs/2511.02650"}
{"aliases": [], "authors": ["Tim R. Davidson", "Adam Fourney", "Saleema Amershi", "Robert West", "Eric Horvitz", "Ece Kamar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f5490062fd59f70a9060044465ad4d8f5c287d3d88be103a8e9b5eacd29aa40e", "submitted_at": "2025-11-04T16:10:57Z", "updated_at": "2026-08-08T16:42:16Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.02687v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.02687", "identifiers": {"arxiv": "2511.02687"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.02687"}], "name": "The Collaboration Gap: Exploration and Benchmarking of Open-World Agentic Cooperation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-08", "url": "https://arxiv.org/abs/2511.02687"}
{"aliases": [], "authors": ["Xingyao Wang", "Simon Rosenberg", "Juan Michelini", "Calvin Smith", "Hoang Tran", "Engel Nyst", "Rohit Malhotra", "Xuhui Zhou", "Valerie Chen", "Robert Brennan", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "299a66a2927cb22220da86acb7bdb1325e8e410496a9a6d0106bf92a4ed55894", "submitted_at": "2025-11-05T18:16:44Z", "updated_at": "2026-04-22T13:47:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.03690v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.03690", "identifiers": {"arxiv": "2511.03690"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.03690"}], "name": "The OpenHands Software Agent SDK: A Composable and Extensible Foundation for Production Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-22", "url": "https://arxiv.org/abs/2511.03690"}
{"aliases": [], "authors": ["Yujian Liu", "Ze Wang", "Hao Chen", "Ximeng Sun", "Xiaodong Yu", "Jialian Wu", "Jiang Liu", "Emad Barsoum", "Zicheng Liu", "Shiyu Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a91413c1e80869860d411402ab49f6e5968787b3bca83f0a9066a0e9c2f75d40", "submitted_at": "2025-11-06T07:29:02Z", "updated_at": "2025-11-06T07:29:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.04137v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.04137", "identifiers": {"arxiv": "2511.04137"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.04137"}], "name": "Learning from Online Videos at Inference Time for Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "This work retrieves online video tutorials and converts them into structured GUI action demonstrations with short objectives. A two-stage selection process supplies a relevant local trajectory as in-context guidance at each execution step.", "topics": ["video", "memory", "planning", "data-generation"], "updated_at": "2025-11-06", "url": "https://arxiv.org/abs/2511.04137"}
{"aliases": [], "authors": ["Alan Seroul", "Théo Fagnoni", "Inès Adnani", "Dana O. Mohamed", "Phillip Kingston"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b20890a7fe72ff0b18d38b139881a09f4f537040fd42f8cfd062ad944fbc7671", "submitted_at": "2025-11-06T09:35:10Z", "updated_at": "2025-11-06T09:35:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.04220v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.04220", "identifiers": {"arxiv": "2511.04220"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.04220"}], "name": "Opus: A Quantitative Framework for Workflow Evaluation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-06", "url": "https://arxiv.org/abs/2511.04220"}
{"aliases": [], "authors": ["Jian Mu", "Chaoyun Zhang", "Chiming Ni", "Lu Wang", "Bo Qiao", "Kartik Mathur", "Qianhui Wu", "Yuhang Xie", "Xiaojun Ma", "Mengyu Zhou", "Si Qin", "Liqun Li", "Yu Kang", "Minghua Ma", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b063a4d2b3a552aa05a153745978473a7d165f2c6b4b97e4c91eff11b7b0f80", "submitted_at": "2025-11-06T12:19:02Z", "updated_at": "2025-11-10T12:27:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.04307v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.04307", "identifiers": {"arxiv": "2511.04307"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.04307"}], "name": "GUI-360$^\\circ$: A Comprehensive Dataset and Benchmark for Computer-Using Agents", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2025-11-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Collects Windows application trajectories with screenshots and action traces for grounding, screen parsing, and action prediction.", "topics": ["data-generation", "grounding", "evaluation"], "updated_at": "2025-11-10", "url": "https://arxiv.org/abs/2511.04307"}
{"aliases": [], "authors": ["Sinchan Snigdha Rej", "Bimalendu Deb"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1487f811ee1d81fc2a09763ddb9ae68ef3c902448f5efb5c7ce4c061ca5e400b", "submitted_at": "2025-11-06T13:41:37Z", "updated_at": "2025-11-06T13:41:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.04359v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.04359", "identifiers": {"arxiv": "2511.04359"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.04359"}], "name": "Neutral-atom quantum computation using multi-qubit geometric gates via adiabatic passage", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-06", "url": "https://arxiv.org/abs/2511.04359"}
{"aliases": [], "authors": ["Lars Krupp", "Daniel Geißler", "Vishal Banwari", "Paul Lukowicz", "Jakob Karolus"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "015d30321b5dad38ce010402c7df8dfcd9aea3dd1c8859ef24696f5eb3f4437f", "submitted_at": "2025-11-06T15:59:59Z", "updated_at": "2025-11-06T15:59:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.04481v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.04481", "identifiers": {"arxiv": "2511.04481"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.04481"}], "name": "Promoting Sustainable Web Agents: Benchmarking and Estimating Energy Consumption through Empirical and Theoretical Analysis", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-06", "url": "https://arxiv.org/abs/2511.04481"}
{"aliases": [], "authors": ["Yigitcan Kaya", "Anton Landerer", "Stijn Pletinckx", "Michelle Zimmermann", "Christopher Kruegel", "Giovanni Vigna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a20b83b94c5a10e2b6fe5fae26b844eb3f4cf9786b0e2edbf5d0b7d9a5d1e8a", "submitted_at": "2025-11-08T02:02:24Z", "updated_at": "2025-11-08T02:02:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.05797v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.05797", "identifiers": {"arxiv": "2511.05797"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.05797"}], "name": "When AI Meets the Web: Prompt Injection Risks in Third-Party AI Chatbot Plugins", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-08", "url": "https://arxiv.org/abs/2511.05797"}
{"aliases": [], "authors": ["Zhaoyang Wang", "Yiming Liang", "Xuchao Zhang", "Qianhui Wu", "Siwei Han", "Anson Bastos", "Rujia Wang", "Chetan Bansal", "Baolin Peng", "Jianfeng Gao", "Saravan Rajmohan", "Huaxiu Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7420d51203de06ed3afe6f2695b8612ec36dc072765a4f2394f366f9e6bf8838", "submitted_at": "2025-11-08T18:45:33Z", "updated_at": "2026-04-13T06:03:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2511.06101v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.06101", "identifiers": {"arxiv": "2511.06101"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.06101"}], "name": "SynthAgent: Adapting Web Agents with Synthetic Supervision", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SynthAgent, which generates tasks by exploring website elements and refines them when observations reveal conflicts. A later trajectory-refinement stage removes misaligned behavior before fine-tuning agents for the target websites.", "topics": ["data-generation", "imitation-learning", "web-navigation"], "updated_at": "2026-04-13", "url": "https://arxiv.org/abs/2511.06101"}
{"aliases": [], "authors": ["Mingde Xu", "Zhen Yang", "Wenyi Hong", "Lihang Pan", "Xinyue Fan", "Yan Wang", "Xiaotao Gu", "Bin Xu", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cde22b8d0792803657599b0f81729a165190a699a8044039edf86a80518921a3", "submitted_at": "2025-11-09T06:58:52Z", "updated_at": "2025-11-09T06:58:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.06251v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.06251", "identifiers": {"arxiv": "2511.06251"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.06251"}], "name": "WebVIA: A Web-based Vision-Language Agentic Framework for Interactive and Verifiable UI-to-Code Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-09", "url": "https://arxiv.org/abs/2511.06251"}
{"aliases": [], "authors": ["Xiangwu Guo", "Difei Gao", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a5f43873a47e5b47250cedd3d860abe2dc5ec9b85fe657444bf2b7378f3e8f7", "submitted_at": "2025-11-09T15:13:45Z", "updated_at": "2025-11-09T15:13:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.06417v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.06417", "identifiers": {"arxiv": "2511.06417"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.06417"}], "name": "AUTO-Explorer: Automated Data Collection for GUI Agent", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Auto-Explorer, an automated GUI exploration method that records software states for training-data collection. UIXplore evaluates exploration coverage, and collected data is used to fine-tune and test GUI element grounding.", "topics": ["data-generation", "grounding", "evaluation"], "updated_at": "2025-11-09", "url": "https://arxiv.org/abs/2511.06417"}
{"aliases": [], "authors": ["Aarash Feizi", "Shravan Nayak", "Xiangru Jian", "Kevin Qinghong Lin", "Kaixin Li", "Rabiul Awal", "Xing Han Lù", "Johan Obando-Ceron", "Juan A. Rodriguez", "Nicolas Chapados", "David Vazquez", "Adriana Romero-Soriano", "Reihaneh Rabbany", "Perouz Taslakian", "Christopher Pal", "Spandana Gella", "Sai Rajeswar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f2b80f294ddf5e3331dcb47a836864032f88f2fdbe35588bfef000cd9ceddc8", "submitted_at": "2025-11-10T17:35:21Z", "updated_at": "2026-06-09T21:30:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.07332v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.07332", "identifiers": {"arxiv": "2511.07332"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.07332"}], "name": "Grounding Computer Use Agents on Human Demonstrations", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "GroundCUA turns expert desktop demonstrations into densely annotated screenshots and grounding instructions. The GroundNext model family uses this supervision and subsequent reinforcement learning, with both grounding benchmarks and an OSWorld agent configuration used for evaluation.", "topics": ["grounding", "data-generation", "evaluation", "reinforcement-learning"], "updated_at": "2026-06-09", "url": "https://arxiv.org/abs/2511.07332"}
{"aliases": [], "authors": ["Soyeong Jeong", "Aparna Elangovan", "Emine Yilmaz", "Oleg Rokhlenko"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aab4bffe546c0f1caf8b1acd8f760038e0875ec04436de09869eadc220687ca2", "submitted_at": "2025-11-11T14:48:34Z", "updated_at": "2025-11-11T14:48:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.08319v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.08319", "identifiers": {"arxiv": "2511.08319"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.08319"}], "name": "Adaptive Multi-Agent Response Refinement in Conversational Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-11", "url": "https://arxiv.org/abs/2511.08319"}
{"aliases": [], "authors": ["Zihan Ma", "Dongsheng Zhu", "Shudong Liu", "Taolin Zhang", "Junnan Liu", "Qingqiu Li", "Minnan Luo", "Songyang Zhang", "Kai Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "efae272be5c4b9f87be8bcaddb471c6c5de03a9b94ce89d03aa188e838975919", "submitted_at": "2025-11-11T17:27:27Z", "updated_at": "2025-11-11T17:27:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.08487v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.08487", "identifiers": {"arxiv": "2511.08487"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.08487"}], "name": "How Brittle is Agent Safety? Rethinking Agent Risk under Intent Concealment and Task Complexity", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-11", "url": "https://arxiv.org/abs/2511.08487"}
{"aliases": [], "authors": ["Ziwei Wang", "Leyang Yang", "Xiaoxuan Tang", "Sheng Zhou", "Dajun Chen", "Wei Jiang", "Yong Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0f61f09ce52b1e66fa81479483ea419aaa16e6b27ce331d3c1039927075fd91", "submitted_at": "2025-11-12T09:06:25Z", "updated_at": "2025-11-12T09:06:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.09127v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.09127", "identifiers": {"arxiv": "2511.09127"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.09127"}], "name": "History-Aware Reasoning for GUI Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "HAR-GUI combines episodic reasoning and short-term memory to use earlier interactions when selecting GUI actions. Its training constructs reflective scenarios and correction guidance and applies a hybrid reinforcement-learning reward to the 3B policy.", "topics": ["interaction-history", "memory", "self-reflection", "reinforcement-learning"], "updated_at": "2025-11-12", "url": "https://arxiv.org/abs/2511.09127"}
{"aliases": [], "authors": ["Leyang Yang", "Ziwei Wang", "Xiaoxuan Tang", "Sheng Zhou", "Dajun Chen", "Wei Jiang", "Yong Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc7f5f0fe1ca373272cda4bbeca3a465e1a7eaec829e8a59bbecbe6049dc7b4a", "submitted_at": "2025-11-12T09:49:31Z", "updated_at": "2025-11-12T09:49:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.09157v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.09157", "identifiers": {"arxiv": "2511.09157"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.09157"}], "name": "ProBench: Benchmarking GUI Agents with Accurate Process Information", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "ProBench extends mobile GUI evaluation with tasks whose correctness depends on the interaction process as well as the final state. Its ProcessProvider supplies process information for assessing agent behavior and analyzing failure modes.", "topics": ["evaluation", "mobile-navigation", "observability"], "updated_at": "2025-11-12", "url": "https://arxiv.org/abs/2511.09157"}
{"aliases": [], "authors": ["Kamil Dreczkowski", "Pietro Vitiello", "Vitalis Vosylius", "Edward Johns"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f5efc9e3b2a51330132f887af2af69cb9fc6af535ce0ef6444eaf817ba3f16e5", "submitted_at": "2025-11-13T09:17:58Z", "updated_at": "2025-11-13T09:17:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.10110v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.10110", "identifiers": {"arxiv": "2511.10110"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.10110"}], "name": "Learning a Thousand Tasks in a Day", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-13", "url": "https://arxiv.org/abs/2511.10110"}
{"aliases": [], "authors": ["Yuan Zhao", "Hualei Zhu", "Tingyu Jiang", "Shen Li", "Xiaohang Xu", "Hao Henry Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "101d38ed439c41ebfa826e990697b502b9126728956295d25b0ea9d539ff23e3", "submitted_at": "2025-11-13T03:41:02Z", "updated_at": "2025-11-13T03:41:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.10705v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.10705", "identifiers": {"arxiv": "2511.10705"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.10705"}], "name": "Co-EPG: A Framework for Co-Evolution of Planning and Grounding in Autonomous GUI Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Co-EPG alternates reinforcement learning for a planner with updates to a grounding model using generated training examples. The evolving grounder provides feedback to the planner, and evaluation uses Multimodal Mind2Web and AndroidControl.", "topics": ["grounding", "planning", "reinforcement-learning", "data-generation"], "updated_at": "2025-11-13", "url": "https://arxiv.org/abs/2511.10705"}
{"aliases": [], "authors": ["Chaoyun Zhang", "Liqun Li", "He Huang", "Chiming Ni", "Bo Qiao", "Si Qin", "Yu Kang", "Minghua Ma", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "00993bbf3f7d13cf8545484cff40c2ad8b942a68f1631f2f7a521bbde08e9826", "submitted_at": "2025-11-14T14:05:31Z", "updated_at": "2026-03-01T12:50:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.11332v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T08:01:28Z", "role": "primary", "sha256": "822cec31fa0db7ffa390fda806d88c2f9c0fe4d328742e7008b662916e92373a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2005.07648%2C2206.04615%2C2311.12022%2C2311.16922%2C2502.13130%2C2506.21506%2C2511.11332&max_results=7"}], "id": "arxiv:2511.11332", "identifiers": {"arxiv": "2511.11332"}, "kind": "paper", "last_verified_at": "2026-10-18T08:01:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.11332"}], "name": "UFO3: Weaving the Digital Agent Galaxy", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-01", "url": "https://arxiv.org/abs/2511.11332"}
{"aliases": [], "authors": ["Zengyi Qin", "Jinyuan Chen", "Yunze Man", "Shengcao Cao", "Ziqi Pang", "Zhuoyuan Wang", "Han Fang", "Ling Zhu", "Zixin Xie", "Zibu Wei", "Tianshu Ran", "Haoran Geng", "Ray Pan", "Qizhen Sun", "Zachary Bright", "Yuyang Cai", "Chongye Yang", "Jiace Zhao", "Tianrui Liu", "Han Cao", "Yeyang Zhou", "Rui Wang", "Song Wang", "Xiang Ren", "Bo Zhang", "Yutong Ban", "Pieter Abbeel", "Brian Anthony"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755d41575ae8ee5d07be0022c8f078c742da4628f12ecd1cad984269441021cf", "submitted_at": "2025-11-11T20:03:38Z", "updated_at": "2026-04-02T02:42:36Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2511.11672v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.11672", "identifiers": {"arxiv": "2511.11672"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.11672"}], "name": "OSGym: Scalable OS Infra for Computer Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-11-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "OSGym supplies full operating-system sandboxes with GUIs for parallel computer-use data collection and training. It combines per-replica state management, hardware-aware orchestration, copy-on-write virtual-machine disks, and layered failure recovery.", "topics": ["sandbox", "efficiency", "reinforcement-learning", "data-generation"], "updated_at": "2026-04-02", "url": "https://arxiv.org/abs/2511.11672"}
{"aliases": [], "authors": ["Hiroki Ohta", "Aaron Merlin Müller", "Shunji Tsuchiya"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f632962bf14387f7d24d6224be0127b3d14f8ae2339abe673f1ab89151edcaa", "submitted_at": "2025-11-15T02:38:34Z", "updated_at": "2026-05-04T19:32:12Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2511.12000v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.12000", "identifiers": {"arxiv": "2511.12000"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.12000"}], "name": "Measurement-Based Quantum Computation Using the Spin-1 XXZ Model with Uniaxial Anisotropy", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-04", "url": "https://arxiv.org/abs/2511.12000"}
{"aliases": [], "authors": ["Yuxiang Zhou", "Jichang Li", "Yanhao Zhang", "Haonan Lu", "Guanbin Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ede08df1c10d8eeede43af75d26608b4917dd83becde23f28cd3adf335f79736", "submitted_at": "2025-11-15T15:22:42Z", "updated_at": "2026-03-12T08:45:07Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2511.12254v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2511.12254", "identifiers": {"arxiv": "2511.12254"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.12254"}], "name": "Mobile-Agent-RAG: Driving Smart Multi-Agent Coordination with Contextual Knowledge Empowerment for Long-Horizon Mobile Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-11-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses separate retrieval systems for high-level mobile-task planning and precise interface operations, with a benchmark for extended cross-app tasks.", "topics": ["mobile-navigation", "memory", "long-horizon"], "updated_at": "2026-03-12", "url": "https://arxiv.org/abs/2511.12254"}
{"aliases": [], "authors": ["Genglin Liu", "Shijie Geng", "Sha Li", "Hejie Cui", "Sarah Zhang", "Xin Liu", "Tianyi Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a105701d9265bacd3971f71f34a0e7ed315b6e54e9a7e675588c50713fcc4de", "submitted_at": "2025-11-17T05:38:50Z", "updated_at": "2026-07-22T21:19:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.12997v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.12997", "identifiers": {"arxiv": "2511.12997"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.12997"}], "name": "WebCoach: Self-Evolving Web Agents with Cross-Session Memory Guidance", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebCoach, which condenses navigation logs into persistent episodic memory and retrieves experiences by similarity and recency. Runtime hooks selectively supply advice to browser agents, evaluated on WebVoyager without retraining the underlying models.", "topics": ["memory", "interaction-history", "web-navigation"], "updated_at": "2026-07-22", "url": "https://arxiv.org/abs/2511.12997"}
{"aliases": [], "authors": ["SeokJoo Kwak", "Jihoon Kim", "Boyoun Kim", "Jung Jae Yoon", "Wooseok Jang", "Jeonghoon Hong", "Jaeho Yang", "Yeong-Dae Kwon"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2cdabdd4fc56b95c2b0c801fefd76d4c724aa0a1d7f762623d4314ca9a38e85c", "submitted_at": "2025-11-17T07:38:05Z", "updated_at": "2025-11-17T07:38:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.13087v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2511.13087", "identifiers": {"arxiv": "2511.13087"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.13087"}], "name": "MEGA-GUI: Multi-stage Enhanced Grounding Agents for GUI Elements", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MEGA-GUI, which separates coarse region selection from fine-grained interface-element grounding. Bidirectional zooming and context-aware instruction rewriting coordinate specialized vision-language agents across visual scales.", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-11-17", "url": "https://arxiv.org/abs/2511.13087"}
{"aliases": [], "authors": ["Yuhan Chen", "Yuxuan Liu", "Long Zhang", "Pengzhi Gao", "Jian Luan", "Wei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "364ceefcbb692493f7da110ec8146411c6dadcb8855b75566f08e50913cbb293", "submitted_at": "2025-11-17T07:43:15Z", "updated_at": "2025-11-17T07:43:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.13091v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2511.13091", "identifiers": {"arxiv": "2511.13091"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.13091"}], "name": "STEP: Success-Rate-Aware Trajectory-Efficient Policy Optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-17", "url": "https://arxiv.org/abs/2511.13091"}
{"aliases": [], "authors": ["Fuyao Zhang", "Jiaming Zhang", "Che Wang", "Xiongtao Sun", "Yurong Hao", "Guowei Guan", "Wenjie Li", "Longtao Huang", "Wei Yang Bryan Lim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d918a5754fe907fc273639f3277a3566d649c08da244714d0d7737456bede230", "submitted_at": "2025-11-17T11:08:45Z", "updated_at": "2025-11-17T11:08:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.13248v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2511.13248", "identifiers": {"arxiv": "2511.13248"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.13248"}], "name": "DualTAP: A Dual-Task Adversarial Protector for Mobile MLLM Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-11-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Applies dual-task adversarial perturbations to mobile screenshots to conceal personally identifiable information from untrusted MLLM routers while preserving agent task utility.", "topics": ["safety-security", "mobile-navigation"], "updated_at": "2025-11-17", "url": "https://arxiv.org/abs/2511.13248"}
{"aliases": [], "authors": ["Sechan Lee", "Hyounghun Kim", "Sangdon Park"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4a0294049fe0a9ff5c59f730330394652b921e144899c7b892c1eea0fdcde5a", "submitted_at": "2025-09-26T02:20:46Z", "updated_at": "2026-09-12T20:13:19Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2511.13725v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.13725", "identifiers": {"arxiv": "2511.13725"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.13725"}], "name": "Can We Stop Malicious AI? KILLBENCH: A Benchmark for External AI Kill Switch Feasibility", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-09-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-12", "url": "https://arxiv.org/abs/2511.13725"}
{"aliases": [], "authors": ["N Dinesh Reddy", "Dylan Snyder", "Lona Kiragu", "Mirajul Mohin", "Shahrear Bin Amin", "Sudeep Pillai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e16ef95be92f1c8806f05d2fc638eb7780f95ab69d3887a9a24929220666f647", "submitted_at": "2025-11-18T07:41:02Z", "updated_at": "2025-11-19T20:07:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.14210v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.14210", "identifiers": {"arxiv": "2511.14210"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.14210"}], "name": "Orion: A Unified Visual Agent for Multimodal Perception, Advanced Visual Reasoning and Execution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-19", "url": "https://arxiv.org/abs/2511.14210"}
{"aliases": [], "authors": ["Fan Yang", "Quanting Xie", "Atsunori Moteki", "Shoichi Masui", "Shan Jiang", "Kanji Uchino", "Yonatan Bisk", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8646f87e7763ebbfd40db237c7f6943eb7ac73d259c2c4674b99c40724f5fb63", "submitted_at": "2025-11-18T22:07:30Z", "updated_at": "2025-11-20T16:11:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.14945v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.14945", "identifiers": {"arxiv": "2511.14945"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.14945"}], "name": "Unsupervised Discovery of Long-Term Spatiotemporal Periodic Workflows in Human Activities", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-20", "url": "https://arxiv.org/abs/2511.14945"}
{"aliases": [], "authors": ["Kevin Qinghong Lin", "Siyuan Hu", "Linjie Li", "Zhengyuan Yang", "Lijuan Wang", "Philip Torr", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d63831bc0ff00979615415ce94abc140b646a4375cb33b950c4a7ccf00bf166", "submitted_at": "2025-11-19T16:00:02Z", "updated_at": "2025-11-19T16:00:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.15567v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.15567", "identifiers": {"arxiv": "2511.15567"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.15567"}], "name": "Computer-Use Agents as Judges for Generative User Interface", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "AUI-Gym evaluates generated websites by having computer-use agents attempt tasks whose executability is checked programmatically. A feedback dashboard condenses navigation histories into visual summaries that a coding agent can use to revise an interface.", "topics": ["evaluation", "ui-design", "observability"], "updated_at": "2025-11-19", "url": "https://arxiv.org/abs/2511.15567"}
{"aliases": [], "authors": ["Yikun Wang", "Zuyan Liu", "Ziyi Wang", "Han Hu", "Pengfei Liu", "Yongming Rao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "682477be3363b830e0067537cbcf81c7fdd22b37d077904f2aafd24b83577d92", "submitted_at": "2025-11-19T18:59:22Z", "updated_at": "2025-12-18T16:47:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.15705v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.15705", "identifiers": {"arxiv": "2511.15705"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.15705"}], "name": "GeoVista: Web-Augmented Agentic Visual Reasoning for Geolocalization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-18", "url": "https://arxiv.org/abs/2511.15705"}
{"aliases": [], "authors": ["Sen Chen", "Tong Zhao", "Yi Bin", "Fei Ma", "Wenqi Shao", "Zheng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1679f7b56c7b60f92195e488c887114bcdcce7975a80fafbf2572c2b1f01ce0b", "submitted_at": "2025-11-20T17:43:46Z", "updated_at": "2025-11-20T17:43:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.16590v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.16590", "identifiers": {"arxiv": "2511.16590"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.16590"}], "name": "D-GARA: A Dynamic Benchmarking Framework for GUI Agent Robustness in Real-World Anomalies", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "D-GARA evaluates Android GUI agents under runtime anomalies such as permission dialogs, battery warnings, and update prompts. Its modular environment embeds these interruptions in Android applications to study recovery and robustness.", "topics": ["evaluation", "robustness", "mobile-navigation"], "updated_at": "2025-11-20", "url": "https://arxiv.org/abs/2511.16590"}
{"aliases": [], "authors": ["Sang Truong", "Yuheng Tu", "Michael Hardy", "Anka Reuel", "Zeyu Tang", "Jirayu Burapacheep", "Jonathan Perera", "Chibuike Uwakwe", "Ben Domingue", "Nick Haber", "Sanmi Koyejo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "399c73cd0c0c6678ec2efd67b52a1aee42d3fc1c47780d4930c3318542145d5a", "submitted_at": "2025-11-20T22:49:21Z", "updated_at": "2025-11-20T22:49:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.16842v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.16842", "identifiers": {"arxiv": "2511.16842"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.16842"}], "name": "Fantastic Bugs and Where to Find Them in AI Benchmarks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-20", "url": "https://arxiv.org/abs/2511.16842"}
{"aliases": [], "authors": ["Horia Cristescu", "Charles Park", "Trong Canh Nguyen", "Sergiu Talmacel", "Alexandru-Gabriel Ilie", "Stefan Adam"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c746b329229d90dbe3d04a138f54c8b205b16ec6173cf681d47a319c8f24ddf", "submitted_at": "2025-11-21T10:47:22Z", "updated_at": "2025-11-21T10:47:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.17131v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.17131", "identifiers": {"arxiv": "2511.17131"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.17131"}], "name": "UI-CUBE: Enterprise-Grade Computer Use Agent Benchmarking Beyond Task Accuracy to Operational Reliability", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-11-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-CUBE evaluates 226 tasks ranging from individual controls to copy-paste and enterprise workflows. Interface variations, multiple resolutions, and application-state validation support diagnosis of operational reliability beyond task accuracy alone.", "topics": ["evaluation", "industry", "workflow", "robustness"], "updated_at": "2025-11-21", "url": "https://arxiv.org/abs/2511.17131"}
{"aliases": [], "authors": ["Juntao Cheng", "Wanyue Zhang", "Zhiwei Yu", "Shuo Ren", "Zheqi He", "Shaoxuan Xie", "Guocai Yao", "Jieru Lin", "Börje F. Karlsson", "Jiajun Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1eaa7548454c3a4aba4b192cd5be3513cd18a170a58e35de5e0f6760a7b9ccfe", "submitted_at": "2025-11-20T09:52:20Z", "updated_at": "2026-07-07T04:04:31Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2511.17649v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.17649", "identifiers": {"arxiv": "2511.17649"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.17649"}], "name": "SWITCH: Benchmarking Modeling and Handling of Tangible Interfaces in Long-horizon Embodied Scenarios", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-07", "url": "https://arxiv.org/abs/2511.17649"}
{"aliases": [], "authors": ["Sandro Rama Fiorini", "Leonardo G. Azevedo", "Raphael M. Thiago", "Valesca M. de Sousa", "Anton B. Labate", "Viviane Torres da Silva"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "410869c0c7f3e2dbc950997e9466086dbabd5e74ac51bf10b31738692ff68225", "submitted_at": "2025-11-21T20:47:41Z", "updated_at": "2025-11-21T20:47:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.17775v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.17775", "identifiers": {"arxiv": "2511.17775"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.17775"}], "name": "Episodic Memory in Agentic Frameworks: Suggesting Next Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-21", "url": "https://arxiv.org/abs/2511.17775"}
{"aliases": [], "authors": ["B. Y. Yan", "Chaofan Li", "Hongjin Qian", "Shuqi Lu", "Zheng Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ed3d008702ec93c56550f82e763534b9c35014392369d904a3fad0ee0e44405c", "submitted_at": "2025-11-23T12:29:33Z", "updated_at": "2025-11-23T12:29:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.18423v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2511.18423", "identifiers": {"arxiv": "2511.18423"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.18423"}], "name": "General Agentic Memory Via Deep Research", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-23", "url": "https://arxiv.org/abs/2511.18423"}
{"aliases": [], "authors": ["Aram Vardanyan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "17debfa3b59e5a2fcae950c2659e5851ae101cc58181e0cc4e0320befe9b4cb1", "submitted_at": "2025-11-22T12:18:35Z", "updated_at": "2025-11-22T12:18:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.19477v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.19477", "identifiers": {"arxiv": "2511.19477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.19477"}], "name": "Building Browser Agents: Architecture, Security, and Practical Solutions", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. Broader claims about the safety of general-purpose browsing are the paper's position, not a catalog conclusion.", "published_at": "2025-11-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Reports architectural and security lessons from building a browser agent, including hybrid accessibility-tree and visual context, browser tooling and programmatic constraints. A WebGames evaluation accompanies the discussion of reliability and specialization.", "topics": ["web-navigation", "safety-security", "tool-integration"], "updated_at": "2025-11-22", "url": "https://arxiv.org/abs/2511.19477"}
{"aliases": [], "authors": ["Yang Liu", "Xiaolong Zhong", "Ling Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c1581f7a86cbcd0cd87c75f19a43d9c2b2534aba3ddde1e010dfb2db9b34440", "submitted_at": "2025-11-23T13:00:47Z", "updated_at": "2025-11-23T13:00:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.19496v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.19496", "identifiers": {"arxiv": "2511.19496"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.19496"}], "name": "Xmodel-2.5: 1.3B Data-Efficient Reasoning SLM", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-23", "url": "https://arxiv.org/abs/2511.19496"}
{"aliases": [], "authors": ["Liqin Luo", "Guangyao Chen", "Xiawu Zheng", "Yongxing Dai", "Yixiong Zou", "Yonghong Tian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50c67671308298167793cbd7a11f24796a3e8008e8ea2d7fdb5a98b5c4d7ffa2", "submitted_at": "2025-11-24T03:11:08Z", "updated_at": "2025-11-26T15:38:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.19516v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.19516", "identifiers": {"arxiv": "2511.19516"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.19516"}], "name": "Connecting the Dots: Training-Free Visual Grounding via Agentic Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-26", "url": "https://arxiv.org/abs/2511.19516"}
{"aliases": [], "authors": ["Ahmed Awadallah", "Yash Lara", "Raghav Magazine", "Hussein Mozannar", "Akshay Nambi", "Yash Pandya", "Aravind Rajeswaran", "Corby Rosset", "Alexey Taymanov", "Vibhav Vineet", "Spencer Whitehead", "Andrew Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0e23ba9b31a01956836e67329652882f84fe8ebef1075385a6af5e68cdae936", "submitted_at": "2025-11-24T19:56:28Z", "updated_at": "2025-11-24T19:56:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.19663v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.19663", "identifiers": {"arxiv": "2511.19663"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.19663"}], "name": "Fara-7B: An Efficient Agentic Model for Computer Use", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-11-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces FaraGen synthetic web trajectories, a compact screenshot-based action model, and WebTailBench for underrepresented browser tasks.", "topics": ["data-generation", "efficiency", "web-navigation"], "updated_at": "2025-11-24", "url": "https://arxiv.org/abs/2511.19663"}
{"aliases": [], "authors": ["Marta Sumyk", "Oleksandr Kosovan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff9343bb706687caf2ea813220ee7221819cd5892ba950f3168b91aebe0a74e0", "submitted_at": "2025-11-25T08:40:33Z", "updated_at": "2025-11-25T08:40:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.20067v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.20067", "identifiers": {"arxiv": "2511.20067"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.20067"}], "name": "\"Are We Done Yet?\": A Vision-Based Judge for Autonomous Task Completion of Computer Use Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study tests vision-language models as task-completion judges using screenshots and instructions from 42 built-in macOS applications. It evaluates judgment accuracy on human-labeled tasks and separately studies how judge feedback affects agent task success.", "topics": ["evaluation", "self-reflection"], "updated_at": "2025-11-25", "url": "https://arxiv.org/abs/2511.20067"}
{"aliases": [], "authors": ["Kaiyuan Zhang", "Mark Tenenholtz", "Kyle Polley", "Jerry Ma", "Denis Yarats", "Ninghui Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ffa50cf03aa91fb1ac2be0588b116c5ae28eb71485d468aceaa0c9b65750339", "submitted_at": "2025-11-25T18:28:35Z", "updated_at": "2026-08-12T00:08:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.20597v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.20597", "identifiers": {"arxiv": "2511.20597"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.20597"}], "name": "BrowseSafe: Understanding and Preventing Prompt Injection Within AI Browser Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-12", "url": "https://arxiv.org/abs/2511.20597"}
{"aliases": [], "authors": ["Ziyue Wang", "Yayati Jadhav", "Peter Pak", "Amir Barati Farimani"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9587dcd7f779f3da4f8f8549bed833300d8b5b1f84a524c0fc9c7964c6c14744", "submitted_at": "2025-11-25T18:55:12Z", "updated_at": "2026-03-19T17:40:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2511.20636v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.20636", "identifiers": {"arxiv": "2511.20636"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.20636"}], "name": "Image2Gcode: Image-to-G-code Generation for Additive Manufacturing Using Diffusion-Transformer Model", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-19", "url": "https://arxiv.org/abs/2511.20636"}
{"aliases": [], "authors": ["Karen Ullrich", "Jingtong Su", "Claudia Shi", "Arjun Subramonian", "Amir Bar", "Ivan Evtimov", "Nikolaos Tsilivis", "Randall Balestriero", "Julia Kempe", "Mark Ibrahim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c288e7c42b5eedeea3f27a5622e87981f101f472db05ead6d6bede1371500f0", "submitted_at": "2025-11-25T19:00:22Z", "updated_at": "2025-11-25T19:00:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.20766v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2511.20766", "identifiers": {"arxiv": "2511.20766"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.20766"}], "name": "OpenApps: Simulating Environment Variations to Measure UI-Agent Reliability", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-11-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces OpenApps, a configurable app ecosystem for measuring UI-agent reliability under changes in appearance and content. Its evaluation compares task success and failure behavior across many versions of the same apps.", "topics": ["evaluation", "robustness", "data-generation"], "updated_at": "2025-11-25", "url": "https://arxiv.org/abs/2511.20766"}
{"aliases": [], "authors": ["Jiayuan Zhang", "Kaiquan Chen", "Zhihao Lu", "Enshen Zhou", "Qian Yu", "Jing Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af9da8a618674b5a40f95029c4797b48a850e0a99de827d3e526b21797f8eef4", "submitted_at": "2025-11-26T13:49:39Z", "updated_at": "2025-11-26T13:49:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.21398v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2511.21398", "identifiers": {"arxiv": "2511.21398"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.21398"}], "name": "Prune4Web: DOM Tree Pruning Programming for Web Agent", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Prune4Web, where generated scoring programs filter DOM elements according to decomposed subtasks before grounding. An annotation pipeline and two-turn training procedure coordinate the planner, programmatic filter and grounder.", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2025-11-26", "url": "https://arxiv.org/abs/2511.21398"}
{"aliases": [], "authors": ["Shaona Ghosh", "Barnaby Simkin", "Kyriacos Shiarlis", "Soumili Nandi", "Dan Zhao", "Matthew Fiedler", "Julia Bazinska", "Nikki Pope", "Roopa Prabhu", "Daniel Rohrer", "Michael Demoret", "Bartley Richardson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c7fbf2ca33fd8c84d3f8f30c6866134041c1db1bee04ef1233427efb92001223", "submitted_at": "2025-11-27T00:19:24Z", "updated_at": "2025-11-27T00:19:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.21990v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.21990", "identifiers": {"arxiv": "2511.21990"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.21990"}], "name": "A Safety and Security Framework for Real-World Agentic Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-27", "url": "https://arxiv.org/abs/2511.21990"}
{"aliases": [], "authors": ["Dasheng Bi", "Yubin Hu", "Mohammed N. Nasir"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d442f9547da046b1350d259207d61b0930553e7c6a190fc280fb6f61795e47d", "submitted_at": "2025-11-27T03:51:49Z", "updated_at": "2025-12-02T22:56:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.22074v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2511.22074", "identifiers": {"arxiv": "2511.22074"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2511.22074"}], "name": "Real-Time Procedural Learning From Experience for AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-02", "url": "https://arxiv.org/abs/2511.22074"}
{"aliases": [], "authors": ["Zehao Deng", "Tianjie Ju", "Zheng Wu", "Zhuosheng Zhang", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9acb395dfaad18b84ea4b52d97be194a42eb42d6ac922c55f2179e38cc258a31", "submitted_at": "2025-11-27T09:01:38Z", "updated_at": "2026-03-04T12:15:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.22235v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2511.22235", "identifiers": {"arxiv": "2511.22235"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.22235"}], "name": "Training High-Level Schedulers with Execution-Feedback Reinforcement Learning for Long-Horizon GUI Automation", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "This work trains high-level schedulers with staged reinforcement learning driven by execution feedback. Its CES framework separates a Coordinator for planning and decomposition, a StateTracker for compressed context, and a low-level Executor.", "topics": ["planning", "state-tracking", "reinforcement-learning", "long-horizon"], "updated_at": "2026-03-04", "url": "https://arxiv.org/abs/2511.22235"}
{"aliases": [], "authors": ["Ruihan Chen", "Qiming Li", "Xiaocheng Feng", "Weihong Zhong", "Xiaoliang Yang", "Yuxuan Gu", "Zekun Zhou", "Yunfei Lu", "Haoyu Ren", "Kun Chen", "Dandan Tu", "Bing Qin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf16987178fc24fb483bfa82efecb1924c7528fd16ac94f1144fcc345748850f", "submitted_at": "2025-11-30T06:47:33Z", "updated_at": "2026-04-28T03:32:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.00756v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.00756", "identifiers": {"arxiv": "2512.00756"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.00756"}], "name": "MPR-GUI: Benchmarking and Enhancing Multilingual Perception and Reasoning in GUI Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-11-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "MPR-GUI evaluates GUI perception and reasoning through strictly aligned tasks in six languages. GUI-XLI adjusts non-English hidden representations toward English representations at language-sensitive layers during inference to address the measured cross-language gaps.", "topics": ["evaluation", "screen-understanding", "robustness"], "updated_at": "2026-04-28", "url": "https://arxiv.org/abs/2512.00756"}
{"aliases": [], "authors": ["Neeraj Anand", "Rishabh Jain", "Sohan Patnaik", "Balaji Krishnamurthy", "Mausoom Sarkar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "faad266b89aed9b210a813b69184d7236f9098c0e095cc426402e154cbdd067b", "submitted_at": "2025-11-30T11:32:54Z", "updated_at": "2025-12-11T18:16:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.00846v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.00846", "identifiers": {"arxiv": "2512.00846"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.00846"}], "name": "AFRAgent : An Adaptive Feature Renormalization Based High Resolution Aware GUI agent", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AFRAgent, which uses adaptive feature renormalization to combine low-resolution visual embeddings with high-resolution details. The architecture is evaluated for smartphone GUI automation on Meta-GUI and AITW.", "topics": ["grounding", "mobile-navigation", "efficiency"], "updated_at": "2025-12-11", "url": "https://arxiv.org/abs/2512.00846"}
{"aliases": [], "authors": ["Hyunjun Kim", "Sooyoung Ryu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e445d0ca79f727a6c1d51e83381c3fd1d8e1fddeff7de99fe70ca06bc1bff49", "submitted_at": "2025-12-01T01:18:21Z", "updated_at": "2025-12-01T01:18:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.01174v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2512.01174", "identifiers": {"arxiv": "2512.01174"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.01174"}], "name": "DrawingBench: Evaluating Spatial Reasoning and UI Interaction Capabilities of Large Language Models through Mouse-Based Drawing Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-12-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "DrawingBench evaluates sequences of low-level mouse-based GUI actions through 250 drawing prompts and rule-based criteria. Action inspection and multi-turn external feedback support analysis of spatial reasoning, tool-state management, and refinement.", "topics": ["evaluation", "grounding", "human-in-the-loop", "state-tracking"], "updated_at": "2025-12-01", "url": "https://arxiv.org/abs/2512.01174"}
{"aliases": [], "authors": ["Xurui Zhou", "Gongwei Chen", "Yuquan Xie", "Zaijing Li", "Kaiwen Zhou", "Shuai Wang", "Shuo Yang", "Zhuotao Tian", "Rui Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a6eb58f9b6349ab9cc33444d8ea82c8f17653cc4a7f4b19ffc94e3b4fb894b3", "submitted_at": "2025-12-01T15:06:45Z", "updated_at": "2026-03-08T10:18:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.01763v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.01763", "identifiers": {"arxiv": "2512.01763"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.01763"}], "name": "HiconAgent: History Context-aware Policy Optimization for GUI Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "HiconAgent trains GUI policies to use history through variable-length context sampling and paired policy updates with compressed and complete histories. Its compression retains past actions while removing historical observations, and an alignment loss couples the two branches.", "topics": ["interaction-history", "memory", "efficiency"], "updated_at": "2026-03-08", "url": "https://arxiv.org/abs/2512.01763"}
{"aliases": [], "authors": ["Aiden Yiliu Li", "Bizhi Yu", "Daoan Lei", "Tianhe Ren", "Shilong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4868c3c85545532e9c208623596d3810c85e32e3cf11429a68847ac11c2ba1cb", "submitted_at": "2025-12-01T18:37:19Z", "updated_at": "2025-12-01T18:37:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.01979v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.01979", "identifiers": {"arxiv": "2512.01979"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.01979"}], "name": "Chain-of-Ground: Improving GUI Grounding via Iterative Reasoning and Reference Feedback", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Chain-of-Ground uses iterative visual reasoning and reference feedback to refine GUI localization without additional training. It evaluates on ScreenSpot-Pro and introduces TPanel-UI to probe grounding on industrial control panels with visual distortions.", "topics": ["grounding", "training-free", "robustness", "evaluation"], "updated_at": "2025-12-01", "url": "https://arxiv.org/abs/2512.01979"}
{"aliases": [], "authors": ["Haolong Yan", "Yeqing Shen", "Xin Huang", "Jia Wang", "Kaijun Tan", "Zhixuan Liang", "Hongxin Li", "Zheng Ge", "Osamu Yoshie", "Si Li", "Xiangyu Zhang", "Daxin Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fda844ac6767e6d9dfd16919042469e2070ec7f092e0b01fb5a243ed6eb4c282", "submitted_at": "2025-12-02T05:11:23Z", "updated_at": "2025-12-02T05:11:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.02423v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2512.02423", "identifiers": {"arxiv": "2512.02423"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.02423"}], "name": "GUI Exploration Lab: Enhancing Screen Navigation in Agents via Multi-Turn Reinforcement Learning", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-12-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI Exploration Lab, a simulation engine for composing screens, interface elements, and navigation graphs. It studies how supervised learning, single-turn RL, and multiturn RL affect GUI navigation and exploration.", "topics": ["reinforcement-learning", "evaluation", "data-generation"], "updated_at": "2025-12-02", "url": "https://arxiv.org/abs/2512.02423"}
{"aliases": [], "authors": ["Jinzhe Tan", "Karim Benyekhlef"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d362a839fd1ee917ebdde6fcd794640523e930cd711769154532200903fe7de5", "submitted_at": "2025-11-28T07:26:29Z", "updated_at": "2025-11-28T07:26:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.04105v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.04105", "identifiers": {"arxiv": "2512.04105"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.04105"}], "name": "LegalWebAgent: Empowering Access to Justice via LLM-Based Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-11-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "LegalWebAgent separates user-needs clarification, webpage browsing, and actions such as form completion or scheduling. A 15-task evaluation studies legal-service workflows relevant to Québec civil-law users, using HTML and screenshot observations.", "topics": ["industry", "web-navigation", "human-in-the-loop", "evaluation"], "updated_at": "2025-11-28", "url": "https://arxiv.org/abs/2512.04105"}
{"aliases": [], "authors": ["Shuyuan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbdf29de1c37cf73d6182c9604fe26a3027417668dcc5a976cb0aa08663e8513", "submitted_at": "2025-12-03T22:36:55Z", "updated_at": "2025-12-03T22:36:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.04302v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2512.04302", "identifiers": {"arxiv": "2512.04302"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.04302"}], "name": "Towards better dense rewards in Reinforcement Learning Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-03", "url": "https://arxiv.org/abs/2512.04302"}
{"aliases": [], "authors": ["Xin Liang", "Zhilin Zhang", "Xiang Zhang", "Haoran Su", "Yiwei Xu", "Siqi Sun", "Chenyu You"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf5f98e8bb0a621bc9ca4e648e4aeb680454c30c2dea9f78499b913f369a41ad", "submitted_at": "2025-12-04T07:22:16Z", "updated_at": "2026-08-22T17:47:59Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.04529v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2512.04529", "identifiers": {"arxiv": "2512.04529"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.04529"}], "name": "SlideGen: Collaborative Multimodal Agents for Scientific Slide Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-22", "url": "https://arxiv.org/abs/2512.04529"}
{"aliases": [], "authors": ["Zhiyuan Jiang", "Shenghao Xie", "Wenyi Li", "Wenqiang Zu", "Peihang Li", "Jiahao Qiu", "Siqi Pei", "Lei Ma", "Tiejun Huang", "Mengdi Wang", "Shilong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bac971bd0e7cd946f81e28d5976eb933cbb95b97ca846ee624a0c36b48760de7", "submitted_at": "2025-12-05T18:39:12Z", "updated_at": "2025-12-05T18:39:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.05941v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.05941", "identifiers": {"arxiv": "2512.05941"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.05941"}], "name": "Zoom in, Click out: Unlocking and Evaluating the Potential of Zooming for GUI Grounding", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "ZoomClick uses repeated zooming and context adjustment to improve GUI grounding without training. The paper studies zoom depth, crop sizes, and related choices and introduces GUIZoom-Bench to evaluate how grounding models respond to zoomed views.", "topics": ["grounding", "training-free", "evaluation", "screen-understanding"], "updated_at": "2025-12-05", "url": "https://arxiv.org/abs/2512.05941"}
{"aliases": [], "authors": ["Alberto Avila-Jimenez", "David Barreda", "Sarah-Laurie Evans", "Austin Luchsinger", "Aiden Massie", "Robert Schweller", "Evan Tomai", "Tim Wylie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "315c1529f91779b8a65df4397f6af16d1cf621307c0882fba42b6b96212ef304", "submitted_at": "2025-12-06T21:32:31Z", "updated_at": "2025-12-06T21:32:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.06574v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.06574", "identifiers": {"arxiv": "2512.06574"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.06574"}], "name": "General Computation using Slidable Tiles with Deterministic Global Forces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-06", "url": "https://arxiv.org/abs/2512.06574"}
{"aliases": [], "authors": ["Shuo Li", "Jiajun Sun", "Zhihao Zhang", "Xiaoran Fan", "Senjie Jin", "Hui Li", "Yuming Yang", "Junjie Ye", "Lixing Shen", "Tao Ji", "Tao Gui", "Qi Zhang", "Xuanjing Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43f79e0d70d06dccdcc23ee8e4686d7307b9ea7b5bbec9771635ca4a2419343a", "submitted_at": "2025-12-07T08:33:55Z", "updated_at": "2025-12-07T08:33:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.06726v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}], "id": "arxiv:2512.06726", "identifiers": {"arxiv": "2512.06726"}, "kind": "paper", "last_verified_at": "2026-10-17T07:57:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.06726"}], "name": "The Role of Entropy in Visual Grounding: Analysis and Optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-07", "url": "https://arxiv.org/abs/2512.06726"}
{"aliases": [], "authors": ["Yulin Li", "Haokun Gui", "Ziyang Fan", "Junjie Wang", "Bin Kang", "Bin Chen", "Zhuotao Tian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a60acb1ca32b554045f00c9f27c70213210f5c32e17da81881c83516703382df", "submitted_at": "2025-12-07T14:42:10Z", "updated_at": "2025-12-07T14:42:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.06866v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.06866", "identifiers": {"arxiv": "2512.06866"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.06866"}], "name": "Less Is More, but Where? Dynamic Token Compression via LLM-Guided Keyframe Prior", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-07", "url": "https://arxiv.org/abs/2512.06866"}
{"aliases": [], "authors": ["Sijia Li", "Yuchen Huang", "Zifan Liu", "Zijian Li", "Jingjing fu", "Lei Song", "Jiang Bian", "Jun Zhang", "Rui Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "130b05efdacec997dc634137dc1cbe36a51876000f04922ac6424ae0b00aa25e", "submitted_at": "2025-12-08T08:27:24Z", "updated_at": "2026-06-28T11:58:42Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.07287v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.07287", "identifiers": {"arxiv": "2512.07287"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.07287"}], "name": "Experience-Evolving Multi-Turn Tool-Use Agent with Hybrid Episodic-Procedural Memory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-28", "url": "https://arxiv.org/abs/2512.07287"}
{"aliases": [], "authors": ["Alisha Ukani", "Hamed Haddadi", "Ali Shahin Shamsabadi", "Peter Snyder"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2576701bbdc4fd8f1f571ed691b06185d1e16bf4955a5c99b10200d796ba6525", "submitted_at": "2025-12-08T17:16:12Z", "updated_at": "2025-12-08T17:16:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.07725v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.07725", "identifiers": {"arxiv": "2512.07725"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.07725"}], "name": "Privacy Practices of Browser Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The abstract describes artifact release as planned; the annotation does not independently verify availability.", "published_at": "2025-12-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents a privacy evaluation framework for browser agents covering component weaknesses, website defenses, tracking, privacy-related prompts and personal-data leakage. The framework is applied to deployed browser-agent tools.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2025-12-08", "url": "https://arxiv.org/abs/2512.07725"}
{"aliases": [], "authors": ["Yinan Zhong", "Qianhao Miao", "Yanjiao Chen", "Jiangyi Deng", "Yushi Cheng", "Wenyuan Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e5275a933450e7cde3f76dd1b32bb562e75dd387b3a0ab5f5972b7b2159d0bc5", "submitted_at": "2025-12-09T09:44:13Z", "updated_at": "2026-08-12T03:36:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.08417v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2512.08417", "identifiers": {"arxiv": "2512.08417"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.08417"}], "name": "Attention is All You Need to Defend Against Indirect Prompt Injection Attacks in LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-12", "url": "https://arxiv.org/abs/2512.08417"}
{"aliases": [], "authors": ["Yunzhu Zhang", "Zeyu Pan", "Zhengwen Zeng", "Shuheng Shen", "Changhua Meng", "Linchao Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95f6e22cda1084d32bbd3f102ea780bfbf392ee578b7c20542314ef76618a56b", "submitted_at": "2025-12-09T12:19:00Z", "updated_at": "2025-12-09T12:19:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.08529v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.08529", "identifiers": {"arxiv": "2512.08529"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.08529"}], "name": "MVP: Multiple View Prediction Improves GUI Grounding", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "MVP addresses unstable GUI coordinates by predicting across several attention-guided cropped views. It clusters the resulting coordinates and selects the densest cluster centroid, evaluating this training-free inference method across grounding models.", "topics": ["grounding", "training-free", "robustness"], "updated_at": "2025-12-09", "url": "https://arxiv.org/abs/2512.08529"}
{"aliases": [], "authors": ["Eranga Bandara", "Ross Gore", "Peter Foytik", "Sachin Shetty", "Ravi Mukkamala", "Abdul Rahman", "Xueping Liang", "Safdar H. Bouk", "Amin Hass", "Sachini Rajapakse", "Ng Wee Keong", "Kasun De Zoysa", "Aruna Withanage", "Nilaan Loganathan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f02c9105f3c92327d8bd17674db28d9082e8ddd3f1a59694565954831bdf162d", "submitted_at": "2025-12-09T16:23:05Z", "updated_at": "2025-12-09T16:23:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.08769v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.08769", "identifiers": {"arxiv": "2512.08769"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.08769"}], "name": "A Practical Guide for Designing, Developing, and Deploying Production-Grade Agentic AI Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-09", "url": "https://arxiv.org/abs/2512.08769"}
{"aliases": [], "authors": ["Zishu Wei", "Qixiang Ma", "Xavier Hu", "Yuhang Liu", "Hui Zang", "Yudong Zhao", "Tao Wang", "Shengyu Zhang", "Fei Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49e76d19d0477bcc56f19e8170fed7e9cc80f9c4bafb5f03af3a2d172cb113bc", "submitted_at": "2025-12-10T07:40:23Z", "updated_at": "2025-12-10T07:40:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.09396v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2512.09396", "identifiers": {"arxiv": "2512.09396"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.09396"}], "name": "GAIR: GUI Automation via Information-Joint Reasoning and Group Reflection", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-12-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GAIR, which combines information from specialized GUI models through a general-purpose coordinating model. When the available information is insufficient, group reflection sends targeted requests back to the specialists before selecting an operation.", "topics": ["planning", "self-reflection", "automation"], "updated_at": "2025-12-10", "url": "https://arxiv.org/abs/2512.09396"}
{"aliases": [], "authors": ["Shizuo Tian", "Hao Wen", "Yuxuan Chen", "Jiacheng Liu", "Shanhui Zhao", "Guohong Liu", "Ju Ren", "Yunxin Liu", "Yuanchun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "72b2db237a52ae31eaba58183a2777e79379aa8e849d18de5d8633daf27ccbf9", "submitted_at": "2025-12-11T07:37:38Z", "updated_at": "2026-05-08T09:40:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.10371v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.10371", "identifiers": {"arxiv": "2512.10371"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.10371"}], "name": "AgentProg: Empowering Long-Horizon GUI Agents with Program-Guided Context Management", "notes": "Reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2025-12-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AgentProg, organizing mobile-agent interaction history as a program with variables and control flow to decide what context to retain. A global belief state handles partial observability, with evaluation on AndroidWorld and extended long-horizon tasks.", "topics": ["interaction-history", "memory", "long-horizon"], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2512.10371"}
{"aliases": [], "authors": ["Yifei He", "Pranit Chawla", "Yaser Souri", "Subhojit Som", "Xia Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7ccc7923f4e6f2ea296b4a6bfc28ba4dc7a52ed256008f153f13fccac3f018f", "submitted_at": "2025-11-22T23:12:56Z", "updated_at": "2026-02-04T19:36:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.10962v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.10962", "identifiers": {"arxiv": "2512.10962"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.10962"}], "name": "WebSTAR: Scalable Data Synthesis for Computer Use Agents with Step-Level Filtering", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-11-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a pipeline that filters individual steps from noisy computer-agent rollouts and augments them with reasoning for training. WebSTAR and WebSCORE support GUI-policy supervision and a distilled process reward model, StepRM.", "topics": ["data-generation", "reward-models", "imitation-learning"], "updated_at": "2026-02-04", "url": "https://arxiv.org/abs/2512.10962"}
{"aliases": [], "authors": ["Frank Zhigang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f0257cc82d6b21083eab757cbb695a505797c369bc98fbd5f81e3ea9f3425d9", "submitted_at": "2025-12-10T22:45:27Z", "updated_at": "2025-12-10T22:45:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.11002v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.11002", "identifiers": {"arxiv": "2512.11002"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.11002"}], "name": "Beyond Memristor: Neuromorphic Computing Using Meminductor", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-10", "url": "https://arxiv.org/abs/2512.11002"}
{"aliases": [], "authors": ["Chunyi Li", "Longfei Li", "Zicheng Zhang", "Xiaohong Liu", "Min Tang", "Weisi Lin", "Guangtao Zhai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "484a51656db4466d5e8630d64fa78d3cf10043568488f5a2e9bae679fa800cbe", "submitted_at": "2025-12-12T14:49:32Z", "updated_at": "2025-12-12T14:49:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.11611v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.11611", "identifiers": {"arxiv": "2512.11611"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.11611"}], "name": "Using GUI Agent for Electronic Design Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-12-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-EDA collects more than 2,000 screenshot-answer-action pairs from engineers working across five CAD tools and five physical domains. The accompanying benchmark evaluates more than 30 GUI agents on specialized electronic-design workflows.", "topics": ["industry", "data-generation", "evaluation", "workflow"], "updated_at": "2025-12-12", "url": "https://arxiv.org/abs/2512.11611"}
{"aliases": [], "authors": ["Luoxi Meng", "Henry Feng", "Ilia Shumailov", "Earlence Fernandes"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa6b4d9fa34610ec6fa8c0ac8b075127365aad57a20d814a2e838a223cf8bd56", "submitted_at": "2025-12-14T08:25:31Z", "updated_at": "2026-03-04T01:49:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.12594v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.12594", "identifiers": {"arxiv": "2512.12594"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.12594"}], "name": "ceLLMate: Sandboxing Browser AI Agents", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2025-12-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "ceLLMate constrains browser-agent authority through an extension that applies policies at the HTTP layer. It evaluates this sandbox against prompt-injection attacks in WASP, avoiding dependence on brittle low-level click and keystroke policies.", "topics": ["safety-security", "sandbox", "web-navigation", "robustness", "evaluation"], "updated_at": "2026-03-04", "url": "https://arxiv.org/abs/2512.12594"}
{"aliases": [], "authors": ["Youngmin Im", "Byeongung Jo", "Jaeyoung Wi", "Seungwoo Baek", "Tae Hoon Min", "Joo Hyung Lee", "Sangeun Oh", "Insik Shin", "Sunjae Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5982901598e43dcadf63b4a950b2fff68e85be71d0f483648b535d82d7745bd7", "submitted_at": "2025-12-14T10:41:39Z", "updated_at": "2026-05-31T23:48:10Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2512.12634v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.12634", "identifiers": {"arxiv": "2512.12634"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.12634"}], "name": "MobiBench: Multi-Branch, Modular Benchmark for Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-12-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobiBench evaluates mobile GUI agents offline while allowing multiple valid action paths. It decomposes agent pipelines into modules to study component contributions, configurations, and performance bottlenecks.", "topics": ["evaluation", "mobile-navigation", "observability"], "updated_at": "2026-05-31", "url": "https://arxiv.org/abs/2512.12634"}
{"aliases": [], "authors": ["Mahir Labib Dihan", "Tanzima Hashem", "Mohammed Eunus Ali", "Md Rizwan Parvez"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "44caacf6646fdbc53e36b4aef93c583e8fae6589109b525816fa12e30ba0b6e3", "submitted_at": "2025-12-14T13:56:54Z", "updated_at": "2025-12-14T13:56:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.12692v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2512.12692", "identifiers": {"arxiv": "2512.12692"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.12692"}], "name": "WebOperator: Action-Aware Tree Search for Autonomous Agents in Web Environment", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-14", "url": "https://arxiv.org/abs/2512.12692"}
{"aliases": [], "authors": ["Jingzhe Ding", "Shengda Long", "Changxin Pu", "Huan Zhou", "Hongwan Gao", "Xiang Gao", "Chao He", "Yue Hou", "Fei Hu", "Zhaojian Li", "Weiran Shi", "Zaiyuan Wang", "Daoguang Zan", "Chenchen Zhang", "Xiaoxu Zhang", "Qizhi Chen", "Xianfu Cheng", "Bo Deng", "Qingshui Gu", "Kai Hua", "Juntao Lin", "Pai Liu", "Mingchen Li", "Xuanguang Pan", "Zifan Peng", "Yujia Qin", "Yong Shan", "Zhewen Tan", "Weihao Xie", "Zihan Wang", "Yishuo Yuan", "Jiayu Zhang", "Enduo Zhao", "Yunfei Zhao", "He Zhu", "Liya Zhu", "Chenyang Zou", "Ming Ding", "Jianpeng Jiao", "Jiaheng Liu", "Minghao Liu", "Qian Liu", "Chongyang Tao", "Jian Yang", "Tong Yang", "Zhaoxiang Zhang", "Xinjie Chen", "Wenhao Huang", "Ge Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a484bbe4f60a828dd51d63b2ce7c4788449c40d90593f8169f86ecb083a3bdc2", "submitted_at": "2025-12-14T15:12:13Z", "updated_at": "2026-01-08T02:42:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.12730v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.12730", "identifiers": {"arxiv": "2512.12730"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.12730"}], "name": "NL2Repo-Bench: Towards Long-Horizon Repository Generation Evaluation of Coding Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-08", "url": "https://arxiv.org/abs/2512.12730"}
{"aliases": [], "authors": ["Chris Latimer", "Nicoló Boschi", "Andrew Neeser", "Chris Bartholomew", "Gaurav Srivastava", "Xuan Wang", "Naren Ramakrishnan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "029574328c1a3d854d03651438e8add96ec40f74cce251c39b3fef84d7a07b3c", "submitted_at": "2025-12-14T19:47:23Z", "updated_at": "2025-12-14T19:47:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.12818v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2512.12818", "identifiers": {"arxiv": "2512.12818"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.12818"}], "name": "Hindsight is 20/20: Building Agent Memory that Retains, Recalls, and Reflects", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-14", "url": "https://arxiv.org/abs/2512.12818"}
{"aliases": [], "authors": ["Emre Can Acikgoz", "Jinoh Oh", "Jie Hao", "Joo Hyuk Jeon", "Heng Ji", "Dilek Hakkani-Tür", "Gokhan Tur", "Xiang Li", "Chengyuan Ma", "Xing Fan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da38bae763c00a56dadfafcd412118bbf3fab525f42eeb2d7ecfada4be181dd1", "submitted_at": "2025-12-15T10:08:53Z", "updated_at": "2025-12-15T10:08:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.13159v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.13159", "identifiers": {"arxiv": "2512.13159"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.13159"}], "name": "SpeakRL: Synergizing Reasoning, Speaking, and Acting in Language Models with Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-15", "url": "https://arxiv.org/abs/2512.13159"}
{"aliases": [], "authors": ["Yuyang Hu", "Shichun Liu", "Yanwei Yue", "Guibin Zhang", "Boyang Liu", "Fangyi Zhu", "Jiahang Lin", "Honglin Guo", "Shihan Dou", "Zhiheng Xi", "Senjie Jin", "Jiejun Tan", "Yanbin Yin", "Jiongnan Liu", "Zeyu Zhang", "Zhongxiang Sun", "Yutao Zhu", "Hao Sun", "Boci Peng", "Zhenrong Cheng", "Xuanbo Fan", "Jiaxin Guo", "Xinlei Yu", "Zhenhong Zhou", "Zewen Hu", "Jiahao Huo", "Junhao Wang", "Yuwei Niu", "Yu Wang", "Zhenfei Yin", "Xiaobin Hu", "Yue Liao", "Qiankun Li", "Kun Wang", "Wangchunshu Zhou", "Yixin Liu", "Dawei Cheng", "Qi Zhang", "Tao Gui", "Shirui Pan", "Yan Zhang", "Philip Torr", "Zhicheng Dou", "Ji-Rong Wen", "Xuanjing Huang", "Yu-Gang Jiang", "Shuicheng Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5b6ea4aa5a4b7f8406226b05f384a40d60ecf3ce4774a3dd958fdc3c1a50e701", "submitted_at": "2025-12-15T17:22:34Z", "updated_at": "2026-01-13T09:33:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.13564v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2512.13564", "identifiers": {"arxiv": "2512.13564"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.13564"}], "name": "Memory in the Age of AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-13", "url": "https://arxiv.org/abs/2512.13564"}
{"aliases": [], "authors": ["Ruanqianqian Huang", "Avery Reyna", "Sorin Lerner", "Haijun Xia", "Brian Hempel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a72fb14063d35cd99805fb4add23272e427ce27bc89df8f26e992192bc84be8", "submitted_at": "2025-12-16T02:15:06Z", "updated_at": "2026-08-18T18:41:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.14012v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.14012", "identifiers": {"arxiv": "2512.14012"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.14012"}], "name": "Professional Software Developers Don't Vibe, They Control: AI Agent Use for Coding in 2025", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-18", "url": "https://arxiv.org/abs/2512.14012"}
{"aliases": [], "authors": ["Shufan Li", "Konstantinos Kallidromitis", "Akash Gokul", "Yusuke Kato", "Kazuki Kozuka", "Aditya Grover"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6228822cf64c6aeb1836bee52a6515123bd0d44d77cceb23e785bb4ff9aef47", "submitted_at": "2025-12-16T02:16:42Z", "updated_at": "2026-10-07T23:06:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.14014v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2512.14014", "identifiers": {"arxiv": "2512.14014"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.14014"}], "name": "MobileWorldBench: Towards Semantic World Modeling For Mobile Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-12-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates semantic world models for mobile interfaces and studies natural-language state transitions as a tool for agent planning.", "topics": ["world-models", "planning", "evaluation"], "updated_at": "2026-10-07", "url": "https://arxiv.org/abs/2512.14014"}
{"aliases": [], "authors": ["Haolong Yan", "Jia Wang", "Xin Huang", "Yeqing Shen", "Ziyang Meng", "Zhimin Fan", "Kaijun Tan", "Jin Gao", "Lieyu Shi", "Mi Yang", "Shiliang Yang", "Zhirui Wang", "Brian Li", "Kang An", "Chenyang Li", "Lei Lei", "Mengmeng Duan", "Danxun Liang", "Guodong Liu", "Hang Cheng", "Hao Wu", "Jie Dong", "Junhao Huang", "Mei Chen", "Renjie Yu", "Shunshan Li", "Xu Zhou", "Yiting Dai", "Yineng Deng", "Yingdan Liang", "Zelin Chen", "Wen Sun", "Chengxu Yan", "Chunqin Xu", "Dong Li", "Fengqiong Xiao", "Guanghao Fan", "Guopeng Li", "Guozhen Peng", "Hongbing Li", "Hang Li", "Hongming Chen", "Jingjing Xie", "Jianyong Li", "Jingyang Zhang", "Jiaju Ren", "Jiayu Yuan", "Jianpeng Yin", "Kai Cao", "Liang Zhao", "Liguo Tan", "Liying Shi", "Mengqiang Ren", "Min Xu", "Manjiao Liu", "Mao Luo", "Mingxin Wan", "Na Wang", "Nan Wu", "Ning Wang", "Peiyao Ma", "Qingzhou Zhang", "Qiao Wang", "Qinlin Zeng", "Qiong Gao", "Qiongyao Li", "Shangwu Zhong", "Shuli Gao", "Shaofan Liu", "Shisi Gao", "Shuang Luo", "Xingbin Liu", "Xiaojia Liu", "Xiaojie Hou", "Xin Liu", "Xuanti Feng", "Xuedan Cai", "Xuan Wen", "Xianwei Zhu", "Xin Liang", "Xin Liu", "Xin Zhou", "Yifan Sui", "Yingxiu Zhao", "Yukang Shi", "Yunfang Xu", "Yuqing Zeng", "Yixun Zhang", "Zejia Weng", "Zhonghao Yan", "Zhiguo Huang", "Zhuoyu Wang", "Zihan Yan", "Zheng Ge", "Jing Li", "Yibo Zhu", "Binxing Jiao", "Xiangyu Zhang", "Daxin Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2fc824bdeabca3451fd895447a0321c507da1ac7bc7cc24184c7768992549daf", "submitted_at": "2025-12-17T13:26:30Z", "updated_at": "2025-12-19T17:36:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.15431v2"}}, "environments": ["android", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2512.15431", "identifiers": {"arxiv": "2512.15431"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.15431"}], "name": "Step-GUI Technical Report", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Step-GUI uses calibrated trajectory rewards to turn generated GUI interaction data into training signals for compact agent models. The report also introduces GUI-MCP for hierarchical GUI operations and local task delegation, and AndroidDaily for static-action and end-to-end mobile evaluation.", "topics": ["data-generation", "reward-models", "tool-integration", "evaluation", "mobile-navigation"], "updated_at": "2025-12-19", "url": "https://arxiv.org/abs/2512.15431"}
{"aliases": [], "authors": ["Nobumasa Ishida", "Yoshihiko Hasegawa"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2ce0ff5029aced93b1fbf1fdad4e1b9068f24f869a4b758e3fe80c67e7bf352", "submitted_at": "2025-12-17T14:16:05Z", "updated_at": "2025-12-17T14:16:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.15472v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.15472", "identifiers": {"arxiv": "2512.15472"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.15472"}], "name": "Energy Inference of Black-Box Quantum Computers Using Quantum Speed Limit", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-17", "url": "https://arxiv.org/abs/2512.15472"}
{"aliases": [], "authors": ["Zhenyu Wu", "Jingjing Xie", "Zehao Li", "Bowen Yang", "Qiushi Sun", "Zhaoyang Liu", "Zhoumianze Liu", "Yu Qiao", "Xiangyu Yue", "Zun Wang", "Zichen Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67a1bd40b117b34e5fb994b1c2c93cd285714097d1bc9730baeeaba6c31a7b8a", "submitted_at": "2025-12-18T08:29:50Z", "updated_at": "2025-12-18T08:29:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.16295v1"}}, "environments": ["web", "desktop", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2512.16295", "identifiers": {"arxiv": "2512.16295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.16295"}], "name": "OS-Oracle: A Comprehensive Framework for Cross-Platform GUI Critic Models", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "OS-Oracle synthesizes GUI critic examples and trains a pre-action critic with supervised learning followed by consistency-preserving GRPO. It introduces OS-Critic Bench across mobile, web, and desktop settings and tests critic-assisted agents in OSWorld and AndroidWorld.", "topics": ["reward-models", "reinforcement-learning", "data-generation", "evaluation"], "updated_at": "2025-12-18", "url": "https://arxiv.org/abs/2512.16295"}
{"aliases": [], "authors": ["Beitong Zhou", "Zhexiao Huang", "Yuan Guo", "Zhangxuan Gu", "Tianyu Xia", "Zichen Luo", "Fei Tang", "Dehan Kong", "Yanyi Shang", "Suling Ou", "Zhenlin Guo", "Changhua Meng", "Shuheng Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4bc8a1da3993ead9b36af47a9332fe9a1a2427101ad0d91da654b4a4b25141df", "submitted_at": "2025-12-18T13:09:09Z", "updated_at": "2025-12-18T13:09:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.16501v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2512.16501", "identifiers": {"arxiv": "2512.16501"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.16501"}], "name": "VenusBench-GD: A Comprehensive Multi-Platform GUI Benchmark for Diverse Grounding Tasks", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "VenusBench-GD is a bilingual, cross-platform GUI grounding benchmark organized into basic and advanced tasks. Its annotation pipeline and six-subtask taxonomy support comparisons of general multimodal models and GUI-specialized models across different grounding demands.", "topics": ["grounding", "evaluation", "data-generation", "robustness"], "updated_at": "2025-12-18", "url": "https://arxiv.org/abs/2512.16501"}
{"aliases": [], "authors": ["SunYoung Park", "Jong-Hyeon Lee", "Youngjune Kim", "Daegyu Sung", "Younghyun Yu", "Young-rok Cha", "Jeongho Ju"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "99991c3c8f776344c1ca4d023e7e9d5db61e8bcc4cf6b86eb23ea9343efeef8d", "submitted_at": "2025-11-04T07:24:45Z", "updated_at": "2026-01-07T06:16:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.16925v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.16925", "identifiers": {"arxiv": "2512.16925"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.16925"}], "name": "V-Agent: An Interactive Video Search System Using Vision-Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-07", "url": "https://arxiv.org/abs/2512.16925"}
{"aliases": [], "authors": ["Saksham Sahai Srivastava", "Haoyu He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cd2c2455fc5ff9eb74efe529fa087c734465646609c5d1cbd8c69b8e9af6c455", "submitted_at": "2025-12-18T08:34:40Z", "updated_at": "2025-12-18T08:34:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.16962v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.16962", "identifiers": {"arxiv": "2512.16962"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.16962"}], "name": "MemoryGraft: Persistent Compromise of LLM Agents via Poisoned Experience Retrieval", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-18", "url": "https://arxiv.org/abs/2512.16962"}
{"aliases": [], "authors": ["Haotian Ye", "Qiyuan He", "Jiaqi Han", "Puheng Li", "Jiaojiao Fan", "Zekun Hao", "Fitsum Reda", "Yogesh Balaji", "Huayu Chen", "Sheng Liu", "Angela Yao", "James Zou", "Stefano Ermon", "Haoxiang Wang", "Ming-Yu Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6fbc1114f4d2c4ea5d9c6379d2140b7fbce656b3e45003b9b75dbb5522c82756", "submitted_at": "2025-12-18T17:13:59Z", "updated_at": "2026-03-22T08:01:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.16975v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.16975", "identifiers": {"arxiv": "2512.16975"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.16975"}], "name": "InfoTok: Adaptive Discrete Video Tokenizer via Information-Theoretic Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-22", "url": "https://arxiv.org/abs/2512.16975"}
{"aliases": [], "authors": ["Brandon Huang", "Hang Hua", "Zhuoran Yu", "Trevor Darrell", "Rogerio Feris", "Roei Herzig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37cd0aa99ed8bad167e25f6e2b48e7cd1cf01249508555f657215213ee94a301", "submitted_at": "2025-12-19T04:09:24Z", "updated_at": "2025-12-31T17:30:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.17221v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.17221", "identifiers": {"arxiv": "2512.17221"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.17221"}], "name": "DAVE: A VLM Vision Encoder for Document Understanding and Web Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "DAVE trains a vision encoder for document images and web agents using self-supervised pretraining followed by supervised parsing and localization. It combines decoder-specific encoder merging with feature fusion and evaluates web localization and agent benchmarks alongside document tasks.", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": "2025-12-31", "url": "https://arxiv.org/abs/2512.17221"}
{"aliases": [], "authors": ["Runze Li", "Yuwen Zhai", "Bo Xu", "LiWu Xu", "Nian Shi", "Wei Zhang", "Ran Lin", "Liang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03cdf8514fa9f036b4614a880d842d49bb9cb0a821858b7215784fba209faa62", "submitted_at": "2025-12-22T13:42:18Z", "updated_at": "2026-04-10T06:20:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.19396v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.19396", "identifiers": {"arxiv": "2512.19396"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.19396"}], "name": "EchoTrail-GUI: Building Actionable Memory for GUI Agents via Critic-Guided Self-Exploration", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "EchoTrail-GUI builds a memory of successful GUI trajectories through autonomous exploration filtered by a reward model. It retrieves relevant trajectories as in-context guidance for new tasks and evaluates the approach on AndroidWorld and AndroidLab.", "topics": ["memory", "data-generation", "reward-models", "mobile-navigation"], "updated_at": "2026-04-10", "url": "https://arxiv.org/abs/2512.19396"}
{"aliases": [], "authors": ["Quyu Kong", "Xu Zhang", "Zhenyu Yang", "Nolan Gao", "Chen Liu", "Panrong Tong", "Chenglin Cai", "Hanzhang Zhou", "Jianan Zhang", "Liangyu Chen", "Zhidan Liu", "Steven Hoi", "Yue Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cf3258fd2286b0ee6babb69b9bfc459ecf40536eec498a3af3403d2971791f3d", "submitted_at": "2025-12-22T14:31:28Z", "updated_at": "2025-12-30T07:15:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.19432v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2512.19432", "identifiers": {"arxiv": "2512.19432"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.19432"}], "name": "MobileWorld: Benchmarking Autonomous Mobile Agents in Agent-User Interactive and MCP-Augmented Environments", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-12-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces long mobile workflows that combine agent-user interaction, cross-application tasks, and MCP-based tool access.", "topics": ["evaluation", "tool-integration", "long-horizon"], "updated_at": "2025-12-30", "url": "https://arxiv.org/abs/2512.19432"}
{"aliases": [], "authors": ["Yue Cao", "Yingyao Wang", "Pi Bu", "Jingxuan Xing", "Wei Jiang", "Zekun Zhu", "Junpeng Ma", "Sashuai Zhou", "Tong Lu", "Jun Song", "Yu Cheng", "Yuning Jiang", "Bo Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "597935923ce9e59720771f3d92e921f2842f908ad81b3ab31a87ed9d9f5c5774", "submitted_at": "2025-12-24T17:40:42Z", "updated_at": "2025-12-24T17:40:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.21302v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.21302", "identifiers": {"arxiv": "2512.21302"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.21302"}], "name": "AndroidLens: Long-latency Evaluation with Nested Sub-targets for Android GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-12-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "AndroidLens provides 571 extended mobile tasks in Chinese and English, with multiple valid paths and preserved environmental anomalies. Its dynamic evaluation uses nested milestones and Average Task Progress to measure partial completion of long workflows.", "topics": ["evaluation", "long-horizon", "mobile-navigation", "robustness"], "updated_at": "2025-12-24", "url": "https://arxiv.org/abs/2512.21302"}
{"aliases": [], "authors": ["Titouan Duston", "Shuo Xin", "Yang Sun", "Daoguang Zan", "Aoyan Li", "Shulin Xin", "Kai Shen", "Yixiao Chen", "Qiming Sun", "Ge Zhang", "Jiashuo Liu", "Huan Zhou", "Jingkai Liu", "Zhichen Pu", "Yuanheng Wang", "Bo-Xuan Ge", "Xin Tong", "Fei Ye", "Zhi-Chao Zhao", "Wen-Biao Han", "Zhoujian Cao", "Yueran Zhao", "Weiluo Ren", "Qingshen Long", "Yuxiao Liu", "Anni Huang", "Yidi Du", "Yuanyuan Rong", "Jiahao Peng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "103cdb182d95ca58b66bb934efcd884b8a90be834d267b6b78316349f4dadbf2", "submitted_at": "2025-12-24T08:11:11Z", "updated_at": "2025-12-24T08:11:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.21373v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.21373", "identifiers": {"arxiv": "2512.21373"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.21373"}], "name": "AInsteinBench: Benchmarking Coding Agents on Scientific Repositories", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-24", "url": "https://arxiv.org/abs/2512.21373"}
{"aliases": [], "authors": ["Shizhe He", "Avanika Narayan", "Ishan S. Khare", "Scott W. Linderman", "Christopher Ré", "Dan Biderman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "107c0d85d1c9cad179302c20a5f86a7f2655228f7d02739ea0dab7233dc28651", "submitted_at": "2025-12-25T15:45:31Z", "updated_at": "2025-12-25T15:45:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.21720v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2512.21720", "identifiers": {"arxiv": "2512.21720"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.21720"}], "name": "An Information Theoretic Perspective on Agentic System Design", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-25", "url": "https://arxiv.org/abs/2512.21720"}
{"aliases": [], "authors": ["Sarthak Mehrotra", "Sairam V C Rebbapragada", "Mani Hemanth Reddy Bonthu", "Vineeth N Balasubramanian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "041466768673c1ea49bceda2ad08784c0fa0e896bcee5452a1e41d51ac3b81d8", "submitted_at": "2025-12-26T12:09:15Z", "updated_at": "2025-12-26T12:09:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.22009v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.22009", "identifiers": {"arxiv": "2512.22009"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.22009"}], "name": "iSHIFT: Lightweight Slow-Fast GUI Agent with Adaptive Perception", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-12-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces iSHIFT, which combines latent reasoning with perception-control tokens to adapt GUI inference. The agent switches between detailed regional grounding and faster use of global visual cues.", "topics": ["efficiency", "grounding", "screen-understanding"], "updated_at": "2025-12-26", "url": "https://arxiv.org/abs/2512.22009"}
{"aliases": [], "authors": ["Hanzhang Zhou", "Xu Zhang", "Panrong Tong", "Jianan Zhang", "Liangyu Chen", "Quyu Kong", "Chenglin Cai", "Chen Liu", "Yue Wang", "Jingren Zhou", "Steven Hoi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a89eb3fd173680c127245933d252d4e245b7179d77bca4b463d50800c78f7238", "submitted_at": "2025-12-26T14:51:52Z", "updated_at": "2025-12-26T14:51:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.22047v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.22047", "identifiers": {"arxiv": "2512.22047"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.22047"}], "name": "MAI-UI Technical Report: Real-World Centric Foundation GUI Agents", "notes": "The annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.", "published_at": "2025-12-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "MAI-UI combines GUI navigation, native user interaction, and MCP tool calls in a family of policies trained through generated data and online reinforcement learning. The report also studies device-cloud routing and scaling of parallel environments and interaction budgets, with grounding and Android navigation evaluation.", "topics": ["human-in-the-loop", "tool-integration", "reinforcement-learning", "data-generation", "efficiency"], "updated_at": "2025-12-26", "url": "https://arxiv.org/abs/2512.22047"}
{"aliases": [], "authors": ["Shaofei Cai", "Yulei Qin", "Haojia Lin", "Zihan Xu", "Gang Li", "Yuchen Shi", "Zongyi Li", "Yong Mao", "Siqi Cai", "Xiaoyu Tan", "Yitao Liang", "Ke Li", "Xing Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e101e5b98d02ddb5b48a7c1ff80d34a6a8ef783609995fcdc719fa1706aa2325", "submitted_at": "2025-12-26T14:51:39Z", "updated_at": "2026-01-06T03:31:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.22322v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2512.22322", "identifiers": {"arxiv": "2512.22322"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.22322"}], "name": "SmartSnap: Proactive Evidence Seeking for Self-Verifying Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-06", "url": "https://arxiv.org/abs/2512.22322"}
{"aliases": [], "authors": ["Phil Cuvin", "Hao Zhu", "Diyi Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "454bc17ab0c9107da3caad1bd33bd4bd6f0ae086789a748b34a0e24e5f8c4505", "submitted_at": "2025-12-28T11:55:20Z", "updated_at": "2026-02-06T11:44:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.22894v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.22894", "identifiers": {"arxiv": "2512.22894"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.22894"}], "name": "DECEPTICON: How Dark Patterns Manipulate Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-06", "url": "https://arxiv.org/abs/2512.22894"}
{"aliases": [], "authors": ["Zhengyang Liang", "Yan Shu", "Xiangrui Liu", "Minghao Qin", "Nicu Sebe", "Jiaxin Mao", "Zheng Liu", "Lizi Liao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "07c9818c14aa54422bde58b1bdf6933a8a4885f35dd00855f64e57ef7aea08ec", "submitted_at": "2025-12-28T19:08:27Z", "updated_at": "2026-10-13T17:54:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.23044v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.23044", "identifiers": {"arxiv": "2512.23044"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.23044"}], "name": "Video-Browser: Towards Open-web Agentic Video Browsing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2512.23044"}
{"aliases": [], "authors": ["Karolina Korgul", "Yushi Yang", "Arkadiusz Drohomirecki", "Piotr Błaszczyk", "Will Howard", "Lukas Aichberger", "Chris Russell", "Philip H. S. Torr", "Adam Mahdi", "Adel Bibi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a255cccbe3f16f498918126724c391911e1d745c093189790cbab715fd28a2d2", "submitted_at": "2025-12-29T01:09:10Z", "updated_at": "2026-07-02T16:12:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.23128v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2512.23128", "identifiers": {"arxiv": "2512.23128"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.23128"}], "name": "It's a TRAP! Task-Redirecting Agent Persuasion Benchmark for Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2025-12-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces TRAP, which evaluates task-redirection attacks embedded in realistic web interfaces. Controlled website clones and a modular injection framework support comparisons of persuasion techniques and contextual changes across agents.", "topics": ["safety-security", "evaluation", "ui-design"], "updated_at": "2026-07-02", "url": "https://arxiv.org/abs/2512.23128"}
{"aliases": [], "authors": ["Baixuan Li", "Jialong Wu", "Wenbiao Yin", "Kuan Li", "Zhongwang Zhang", "Huifeng Yin", "Zhengwei Tao", "Liwen Zhang", "Pengjun Xie", "Jingren Zhou", "Yong Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fde71dd935acb953bc5d01bcd48f0d6c3351db6a66126761baf134c114e7893c", "submitted_at": "2025-12-29T17:59:14Z", "updated_at": "2025-12-29T17:59:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.23647v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.23647", "identifiers": {"arxiv": "2512.23647"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.23647"}], "name": "Nested Browser-Use Learning for Agentic Information Seeking", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2025-12-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "NestBrowse organizes browser actions in a nested framework that separates interaction control from page exploration. It evaluates this design for deep information seeking beyond snippet retrieval and direct URL fetching.", "topics": ["web-navigation", "planning", "efficiency", "tool-integration"], "updated_at": "2025-12-29", "url": "https://arxiv.org/abs/2512.23647"}
{"aliases": [], "authors": ["Adharsh Kamath", "Sishen Zhang", "Calvin Xu", "Shubham Ugare", "Gagandeep Singh", "Sasa Misailovic"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c9530d8aa508c434fe055f745193b7f19ffac430bbdce3ae2c4f60c07af8e39", "submitted_at": "2025-12-25T06:12:13Z", "updated_at": "2025-12-25T06:12:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.23738v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2512.23738", "identifiers": {"arxiv": "2512.23738"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2512.23738"}], "name": "Enforcing Temporal Constraints for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-12-25", "url": "https://arxiv.org/abs/2512.23738"}
{"aliases": ["ShowUI-pi", "ShowUI Pi"], "authors": ["Siyuan Hu", "Kevin Qinghong Lin", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b7d1f358b450f2a6c3deb00f1b16c54d443f48fda3a7dc70c77456bebf35ad7c", "submitted_at": "2025-12-31T16:51:14Z", "updated_at": "2025-12-31T16:51:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.24965v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2512.24965", "identifiers": {"arxiv": "2512.24965"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.24965"}], "name": "ShowUI-$π$: Flow-based Generative Models as GUI Dexterous Hands", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-12-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Models discrete clicks and continuous drags with flow-based action generation, and introduces drag trajectories and the ScreenDrag benchmark.", "topics": ["grounding", "evaluation", "data-generation"], "updated_at": "2025-12-31", "url": "https://arxiv.org/abs/2512.24965"}
{"aliases": [], "authors": ["Bin Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c86bf07cf2b3485acfebfe66fc6d72c535da58f02a1359473c314a58fe2bd104", "submitted_at": "2026-01-05T02:38:40Z", "updated_at": "2026-01-05T02:38:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.01743v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.01743", "identifiers": {"arxiv": "2601.01743"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.01743"}], "name": "AI Agent Systems: Architectures, Applications, and Evaluation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-05", "url": "https://arxiv.org/abs/2601.01743"}
{"aliases": [], "authors": ["Shaocheng Shen", "Jianfeng Liang", "Chunlei Cai", "Cong Geng", "Huiyu Duan", "Xiaoyun Zhang", "Qiang Hu", "Guangtao Zhai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "684067a1492e6e4363cb5eedf7931badf7b253852324dd686666e31f5534a32c", "submitted_at": "2026-01-05T12:06:43Z", "updated_at": "2026-03-16T06:59:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.02046v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.02046", "identifiers": {"arxiv": "2601.02046"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.02046"}], "name": "Agentic Retoucher for Text-To-Image Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-16", "url": "https://arxiv.org/abs/2601.02046"}
{"aliases": [], "authors": ["Chuanrui Hu", "Xingze Gao", "Zuyi Zhou", "Dannong Xu", "Yi Bai", "Xintong Li", "Hui Zhang", "Tong Li", "Chong Zhang", "Lidong Bing", "Yafeng Deng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a27e46c1ebad9ae549abfed18965dec3b780281bb6ce77a3301b71700afefea5", "submitted_at": "2026-01-05T14:39:43Z", "updated_at": "2026-01-09T02:23:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.02163v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.02163", "identifiers": {"arxiv": "2601.02163"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.02163"}], "name": "EverMemOS: A Self-Organizing Memory Operating System for Structured Long-Horizon Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-09", "url": "https://arxiv.org/abs/2601.02163"}
{"aliases": [], "authors": ["Samuele Marro", "Alan Chan", "Xinxing Ren", "Lewis Hammond", "Jesse Wright", "Gurjyot Wanga", "Tiziano Piccardi", "Nuno Campos", "Tobin South", "Jialin Yu", "Sunando Sengupta", "Eric Sommerlade", "Alex Pentland", "Philip Torr", "Jiaxin Pei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c106bef05d24b48f12eac4702bf9600aebb5805eff075aac7c33b8056640ce9c", "submitted_at": "2025-12-07T17:45:01Z", "updated_at": "2026-01-12T23:23:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.02371v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.02371", "identifiers": {"arxiv": "2601.02371"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.02371"}], "name": "Permission Manifests for Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-12-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2601.02371"}
{"aliases": [], "authors": ["Jiaxin Ai", "Yukang Feng", "Fanrui Zhang", "Jianwen Sun", "Zizhen Li", "Chuanhao Li", "Yifan Chang", "Wenxiao Wu", "Ruoxi Wang", "Mingliang Zhai", "Kaipeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "afab3dcd2444bf8333f2a2fea887819f6fca55c41209d8c449f6211d1c62b819", "submitted_at": "2025-12-30T01:49:46Z", "updated_at": "2025-12-30T01:49:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.02399v1"}}, "environments": ["windows", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:36Z", "role": "primary", "sha256": "f9a44f8090369a296a584a9b94c4491662ee25bb8f7239ef75dfd8f775be8887", "supports": ["summary", "environments", "scope"], "url": "https://arxiv.org/html/2601.02399v1"}], "id": "arxiv:2601.02399", "identifiers": {"arxiv": "2601.02399"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.02399"}], "name": "ProSoftArena: Benchmarking Hierarchical Capabilities of Multimodal Agents in Professional Software Environments", "notes": "The paper abstract and official full-text sections were reviewed. The full text supports Windows 11 virtual-machine execution, screenshot/state observations, and mouse/keyboard actions. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2025-12-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "ProSoftArena evaluates 436 professional-software tasks across six disciplines and 13 applications in a Windows 11 virtual-machine environment. Its capability hierarchy, execution-based checks, and human-in-the-loop evaluation cover precise GUI operations and multi-software workflows.", "topics": ["evaluation", "industry", "science", "cross-app"], "updated_at": "2025-12-30", "url": "https://arxiv.org/abs/2601.02399"}
{"aliases": [], "authors": ["Hao Bai", "Alexey Taymanov", "Tong Zhang", "Aviral Kumar", "Spencer Whitehead"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae8751c4fae3b4f838eba59ffa8be37a1a5f4fa497dfd69fa3075f8581547f51", "submitted_at": "2026-01-05T09:35:11Z", "updated_at": "2026-05-02T18:45:09Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2601.02439v6"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.02439", "identifiers": {"arxiv": "2601.02439"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.02439"}], "name": "WebGym: Scaling Training Environments for Visual Web Agents with Realistic Tasks", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebGym supplies realistic web tasks with rubric-based evaluation and an asynchronous rollout system for reinforcement learning. It studies how training on an agent’s own interaction traces scales with task coverage and transfers to unseen websites.", "topics": ["web-navigation", "reinforcement-learning", "data-generation", "evaluation", "efficiency"], "updated_at": "2026-05-02", "url": "https://arxiv.org/abs/2601.02439"}
{"aliases": [], "authors": ["Saba Naqvi", "Mohammad Baqar", "Nawaz Ali Mohammad"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "38c501e7383399b0270897fe71ea2ab3c0ad36223613ac7a2bc737cbc4b3373d", "submitted_at": "2026-01-05T18:20:14Z", "updated_at": "2026-01-05T18:20:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.02454v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.02454", "identifiers": {"arxiv": "2601.02454"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.02454"}], "name": "The Rise of Agentic Testing: Multi-Agent Systems for Robust Software Quality Assurance", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-05", "url": "https://arxiv.org/abs/2601.02454"}
{"aliases": [], "authors": ["Theresa Farah", "Loïc Flis", "Pierre Laly", "Guo-En Chang", "Jun-Yu Ou", "Yoshishige Tsuchiya", "Yan Pennec", "Bahram Djafari-Rouhani", "Xin Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d25ebe888d350bc36165e621c7811dea954bcd7bafba96bbfd27985da3bbc89", "submitted_at": "2026-01-06T00:31:47Z", "updated_at": "2026-03-24T10:48:35Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.02617v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.02617", "identifiers": {"arxiv": "2601.02617"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.02617"}], "name": "Implementation of Reservoir Computing Using Coupled Microelectromechanical Drum Resonators via Sideband-Pumped Phonon-Cavity Dynamics", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-24", "url": "https://arxiv.org/abs/2601.02617"}
{"aliases": [], "authors": ["Shuhaib Mehri", "Priyanka Kargupta", "Tal August", "Dilek Hakkani-Tür"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d613c8ba55e91a61afd26e1398b28d99c8ea0786b25ad484c42552423674619b", "submitted_at": "2026-01-06T04:26:22Z", "updated_at": "2026-03-15T04:00:52Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.02702v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.02702", "identifiers": {"arxiv": "2601.02702"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.02702"}], "name": "MultiSessionCollab: Learning User Preferences with Memory to Improve Long-Term Collaboration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-15", "url": "https://arxiv.org/abs/2601.02702"}
{"aliases": [], "authors": ["Yunqi Hong", "Kuei-Chun Kao", "Hengguang Zhou", "Cho-Jui Hsieh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f87082c8bdec097a265a676d4aada136bdd2019bb0c4e7c675e619ad024235b3", "submitted_at": "2026-01-06T23:43:47Z", "updated_at": "2026-08-31T08:33:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.03468v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.03468", "identifiers": {"arxiv": "2601.03468"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.03468"}], "name": "Understanding Reward Hacking in Text-to-Image Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2601.03468"}
{"aliases": [], "authors": ["Zheng Wu", "Xingyu Lou", "Xinbei Ma", "Yansi Li", "Weiwen Liu", "Weinan Zhang", "Jun Wang", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f0592eb3fbfbf8cd2842a6f873ebac8773bf86f66f12dd1376a894f749606df", "submitted_at": "2026-01-07T06:43:50Z", "updated_at": "2026-04-11T14:07:11Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2601.03641v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2601.03641", "identifiers": {"arxiv": "2601.03641"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.03641"}], "name": "Agent-Dice: Disentangling Knowledge Updates via Geometric Consensus for Agent Continual Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-11", "url": "https://arxiv.org/abs/2601.03641"}
{"aliases": [], "authors": ["Mingyu Ouyang", "Kevin Qinghong Lin", "Mike Zheng Shou", "Hwee Tou Ng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "518a9699a6c11e6f61b655f4016170c50bf5d949e477134521df1c0214320046", "submitted_at": "2026-01-07T13:48:12Z", "updated_at": "2026-01-07T13:48:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.03928v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.03928", "identifiers": {"arxiv": "2601.03928"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.03928"}], "name": "FocusUI: Efficient UI Grounding via Position-Preserving Visual Token Selection", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-01-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "FocusUI selects instruction-relevant screenshot patches while preserving their positional structure for GUI grounding. Its PosPad scheme replaces runs of dropped tokens with positional markers and is evaluated under reduced token budgets.", "topics": ["grounding", "efficiency", "screen-understanding"], "updated_at": "2026-01-07", "url": "https://arxiv.org/abs/2601.03928"}
{"aliases": [], "authors": ["Yilin Cao", "Yufeng Zhong", "Zhixiong Zeng", "Siran Dai", "Liming Zheng", "Jing Huang", "Haibo Qiu", "Peng Shi", "Wenji Mao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c40e3db3f538f089c19d7c8233ea01174d0591c3211a7acc412ee4df28e902a8", "submitted_at": "2026-01-07T15:51:44Z", "updated_at": "2026-08-31T13:27:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.04035v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.04035", "identifiers": {"arxiv": "2601.04035"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.04035"}], "name": "MobileDreamer: Generative Sketch World Model for GUI Agent", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileDreamer predicts post-action GUI states as task-relevant textual sketches with an order-invariant learning objective. Agents use these predictions for lookahead action selection, evaluated on AndroidWorld.", "topics": ["world-models", "planning", "mobile-navigation"], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2601.04035"}
{"aliases": [], "authors": ["Ziyun Zhang", "Zezhou Wang", "Xiaoyi Zhang", "Zongyu Guo", "Jiahao Li", "Bin Li", "Yan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4d0a59f1eb8197b4d389f5f40957fb8aeb1e22d0739f9877b26d57ed70fe59b", "submitted_at": "2026-01-07T17:40:08Z", "updated_at": "2026-06-30T13:54:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.04126v3"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.04126", "identifiers": {"arxiv": "2601.04126"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.04126"}], "name": "InfiniteWeb: Scalable Web Environment Synthesis for GUI Agent Training", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "InfiniteWeb generates functional, interconnected websites and task evaluators using shared specifications and task-driven development. It evaluates website construction and the transfer of GUI policies trained in these synthetic web environments to OSWorld and Online-Mind2Web.", "topics": ["data-generation", "reinforcement-learning", "evaluation", "sandbox"], "updated_at": "2026-06-30", "url": "https://arxiv.org/abs/2601.04126"}
{"aliases": [], "authors": ["Yifei Gao", "Jiang Wu", "Xiaoyi Chen", "Yifan Yang", "Zhe Cui", "Tianyi Ma", "Jiaming Zhang", "Jitao Sang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47a340b19cf399f2ae7851482030d85e572b0f6ad8415c67657beec737a80e06", "submitted_at": "2026-01-08T02:07:53Z", "updated_at": "2026-01-08T02:07:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.04500v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.04500", "identifiers": {"arxiv": "2601.04500"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.04500"}], "name": "GUITester: Enabling GUI Agents for Exploratory Defect Discovery", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUITestBench for interactive defect discovery and GUITester, which separates task navigation from anomaly verification. Testing-oriented plans and history-based reflection help distinguish software defects from agent execution mistakes.", "topics": ["automation", "evaluation", "self-reflection"], "updated_at": "2026-01-08", "url": "https://arxiv.org/abs/2601.04500"}
{"aliases": [], "authors": ["Qiang Yu", "Xinran Cheng", "Chuanyi Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b53fc67020db1bbe1ae09a7531bee0aa0638e331a4252da7b9692222bae924c2", "submitted_at": "2026-01-08T10:21:56Z", "updated_at": "2026-01-08T10:21:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.04795v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.04795", "identifiers": {"arxiv": "2601.04795"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.04795"}], "name": "Defense Against Indirect Prompt Injection via Tool Result Parsing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-08", "url": "https://arxiv.org/abs/2601.04795"}
{"aliases": [], "authors": ["Muzhao Tian", "Zisu Huang", "Xiaohua Wang", "Jingwen Xu", "Zhengkang Guo", "Qi Qian", "Yuanzhe Shen", "Kaitao Song", "Jiakang Yuan", "Changze Lv", "Xiaoqing Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e71d3daac265787fac2a63f9786645b8f5acf3959df08bd458409818eba1988e", "submitted_at": "2026-01-08T16:54:30Z", "updated_at": "2026-08-31T14:26:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.05107v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.05107", "identifiers": {"arxiv": "2601.05107"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.05107"}], "name": "Controllable Memory Usage: Balancing Anchoring and Innovation in Long-Term Human-Agent Interaction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2601.05107"}
{"aliases": [], "authors": ["Yuxiang Ji", "Yong Wang", "Ziyu Ma", "Yiming Hu", "Hailang Huang", "Xuecai Hu", "Guanhua Chen", "Liaoni Wu", "Xiangxiang Chu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d66de194aa18bd8fcc88ecd10e3bf5ae71f7f4159f29c51257c21d529bf8f92", "submitted_at": "2026-01-08T23:47:30Z", "updated_at": "2026-01-08T23:47:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.05432v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.05432", "identifiers": {"arxiv": "2601.05432"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.05432"}], "name": "Thinking with Map: Reinforced Parallel Map-Augmented Agent for Geolocalization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-08", "url": "https://arxiv.org/abs/2601.05432"}
{"aliases": [], "authors": ["Zezhou Wang", "Ziyun Zhang", "Xiaoyi Zhang", "Zhuzhong Qian", "Yan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e342723078b50e612df749df80eadfd16e0fe29f5875777c1b3189d8b5981512", "submitted_at": "2026-01-09T13:26:38Z", "updated_at": "2026-02-10T15:12:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.05787v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.05787", "identifiers": {"arxiv": "2601.05787"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.05787"}], "name": "From Off-Policy to On-Policy: Enhancing GUI Agents via Bi-level Expert-to-Policy Assimilation", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "BEPA adapts a small set of expert GUI traces into policy-aligned guidance through trajectories reachable by the learner and a dynamically updated task cache. It evaluates this two-level approach to reinforcement learning on desktop and web benchmarks.", "topics": ["reinforcement-learning", "data-generation", "evaluation"], "updated_at": "2026-02-10", "url": "https://arxiv.org/abs/2601.05787"}
{"aliases": [], "authors": ["Zeyi Liao", "Yadong Lu", "Boyu Gou", "Huan Sun", "Ahmed Awadallah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8e95ae74671c37b630f10acd23cf36ef2fca9570858647a165d78a6f9a4695b", "submitted_at": "2025-11-07T19:40:09Z", "updated_at": "2025-11-07T19:40:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.06031v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.06031", "identifiers": {"arxiv": "2601.06031"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.06031"}], "name": "Beyond Clicking:A Step Towards Generalist GUI Grounding via Text Dragging", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2025-11-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2025-11-07", "url": "https://arxiv.org/abs/2601.06031"}
{"aliases": [], "authors": ["Ziqiao Xi", "Shuang Liang", "Qi Liu", "Jiaqing Zhang", "Letian Peng", "Fang Nan", "Meshal Nayim", "Tianhui Zhang", "Rishika Mundada", "Lianhui Qin", "Biwei Huang", "Kun Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d9cf435c050583a25e743856273f969e5fefd4bc22dc7c2323722f19796e3cd", "submitted_at": "2026-01-09T21:59:31Z", "updated_at": "2026-04-19T04:36:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.06328v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.06328", "identifiers": {"arxiv": "2601.06328"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.06328"}], "name": "C-World: A Computer Use Agent Environment Creator", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-19", "url": "https://arxiv.org/abs/2601.06328"}
{"aliases": [], "authors": ["Zhengxuan Lu", "Dongfang Li", "Yukun Shi", "Beilun Wang", "Longyue Wang", "Baotian Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95fb14d9a3aed69bc3dec08d87eed8be0e22c9c997ea95e486e12f58657ef7ee", "submitted_at": "2026-01-10T03:17:25Z", "updated_at": "2026-02-10T08:51:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.06411v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.06411", "identifiers": {"arxiv": "2601.06411"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.06411"}], "name": "Structured Episodic Event Memory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-10", "url": "https://arxiv.org/abs/2601.06411"}
{"aliases": [], "authors": ["Qiang Zhang", "Boli Chen", "Fanrui Zhang", "Ruixue Ding", "Shihang Wang", "Qiuchen Wang", "Yinfeng Huang", "Haonan Zhang", "Rongxiang Zhu", "Pengyong Wang", "Ailin Ren", "Xin Li", "Pengjun Xie", "Jiawei Liu", "Ning Guo", "Jingren Zhou", "Zheng-Jun Zha"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bda6e70050d578b2eb3e9bb89d801bf8afc916beb2dd8b61db8abd0317ecf7c8", "submitted_at": "2026-01-10T08:43:07Z", "updated_at": "2026-07-22T09:23:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.06487v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.06487", "identifiers": {"arxiv": "2601.06487"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.06487"}], "name": "ArenaRL: Scaling RL for Open-Ended Agents via Tournament-based Relative Ranking", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-22", "url": "https://arxiv.org/abs/2601.06487"}
{"aliases": [], "authors": ["Wenyu Mao", "Haosong Tan", "Shuchang Liu", "Haoyang Liu", "Yifan Xu", "Huaxiang Ji", "Xiang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f291b5dcd2245f9daa734f08d6bbeed105207f58a3b129e5dae89c2b2ef67b8", "submitted_at": "2026-01-10T08:46:53Z", "updated_at": "2026-01-10T08:46:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.06490v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.06490", "identifiers": {"arxiv": "2601.06490"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.06490"}], "name": "Bi-Mem: Bidirectional Construction of Hierarchical Memory for Personalized LLMs via Inductive-Reflective Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-10", "url": "https://arxiv.org/abs/2601.06490"}
{"aliases": [], "authors": ["Pei Yang", "Wanyi Chen", "Ke Wang", "Lynn Ai", "Eric Yang", "Tianyu Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "751737e72e9d2a979957ada00fa0514c1eac97d2ba697f7a0b27fcdbcfcde1c1", "submitted_at": "2026-01-10T13:25:27Z", "updated_at": "2026-04-07T01:56:13Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2601.06565v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.06565", "identifiers": {"arxiv": "2601.06565"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.06565"}], "name": "EVM-QuestBench: An Execution-Grounded Benchmark for Natural-Language Transaction Code Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-07", "url": "https://arxiv.org/abs/2601.06565"}
{"aliases": [], "authors": ["Kaiwen Zhou", "Shreedhar Jangam", "Ashwin Nagarajan", "Tejas Polu", "Suhas Oruganti", "Chengzhi Liu", "Ching-Chen Kuo", "Yuting Zheng", "Sravana Narayanaraju", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7065f89a42957d2d950f174452c7d7ec4ea4a00d03d6decfef97dcab4a676b34", "submitted_at": "2026-01-10T19:53:09Z", "updated_at": "2026-01-13T18:20:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.06663v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.06663", "identifiers": {"arxiv": "2601.06663"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.06663"}], "name": "SafePro: Evaluating the Safety of Professional-Level AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-13", "url": "https://arxiv.org/abs/2601.06663"}
{"aliases": [], "authors": ["Jikai Chen", "Long Chen", "Dong Wang", "Qinglin Su", "Zhixuan Chu", "Bingguang Hao", "Leilei Gan", "Chenyi Zhuang", "Jinjie Gu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd2b645b974322737d5b3398eb3d57bff0c86f0b083bb8202d7bc5c0316ce12b", "submitted_at": "2026-01-11T12:54:54Z", "updated_at": "2026-01-16T06:33:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.06899v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["identifiers.preferred_arxiv", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.06899", "identifiers": {"arxiv": "2601.06899", "preferred_arxiv": "2508.13634"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.06899"}], "name": "V2P: Visual Attention Calibration for GUI Grounding via Background Suppression and Center Peaking", "notes": "The authors identify this arXiv entry as an intended replacement of arXiv:2508.13634 and direct subsequent updates to https://arxiv.org/abs/2508.13634. Retained as a bibliographic provenance record; not counted as an additional reviewed work.", "published_at": "2026-01-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-16", "url": "https://arxiv.org/abs/2601.06899"}
{"aliases": [], "authors": ["Hao Wang", "Wenhui Zhu", "Shao Tang", "Zhipeng Wang", "Xuanzhao Dong", "Xin Li", "Xiwen Chen", "Ashish Bastola", "Xinhao Huang", "Yalin Wang", "Abolfazl Razi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "460abc8cf63d1a113c661cac6db923e62aeaaa438dea96b5fb7180a6d96afc0f", "submitted_at": "2026-01-12T02:19:34Z", "updated_at": "2026-01-12T02:19:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07143v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.07143", "identifiers": {"arxiv": "2601.07143"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.07143"}], "name": "EZBlender: Efficient 3D Editing with Plan-and-ReAct Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2601.07143"}
{"aliases": [], "authors": ["Yichun Zhang", "Xiangwu Guo", "Yauhong Goh", "Jessica Hu", "Zhiheng Chen", "Xin Wang", "Difei Gao", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60c6a513f53c16ea6d989a7b2be56ad26b83e531bcea46b00f2e08476b686c97", "submitted_at": "2026-01-12T04:04:20Z", "updated_at": "2026-01-12T04:04:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07181v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.07181", "identifiers": {"arxiv": "2601.07181"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.07181"}], "name": "ShowUI-Aloha: Human-Taught GUI Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Transforms human screen recordings and input events into annotated demonstrations used by planning and execution components for GUI tasks.", "topics": ["recording", "workflow", "data-generation"], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2601.07181"}
{"aliases": [], "authors": ["Jihong Wang", "Jiamu Zhou", "Weiming Zhang", "Teng Wang", "Weiwen Liu", "Zhuosheng Zhang", "Xingyu Lou", "Weinan Zhang", "Huarong Deng", "Jun Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2afe8d0d8d83f91003b7885635e098e19f69024306c43c3edb23c110e553f63", "submitted_at": "2026-01-12T07:08:42Z", "updated_at": "2026-04-23T09:34:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.07262v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.07262", "identifiers": {"arxiv": "2601.07262"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.07262"}], "name": "ColorBrowserAgent: Complex Long-Horizon Browser Agent with Adaptive Knowledge Evolution", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "ColorBrowserAgent converts human feedback into reusable website knowledge and uses that knowledge to guide progressive context summaries. It evaluates the resulting memory and adaptation mechanisms on long browser workflows.", "topics": ["memory", "human-in-the-loop", "long-horizon", "web-navigation"], "updated_at": "2026-04-23", "url": "https://arxiv.org/abs/2601.07262"}
{"aliases": [], "authors": ["Xinyi Wu", "Geng Hong", "Yueyue Chen", "MingXuan Liu", "Feier Jin", "Xudong Pan", "Jiarun Dai", "Baojun Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4af00e0632ee21f70409998fc22f67e9a4ed1de415b488eda122c6ef94cab7aa", "submitted_at": "2026-01-12T07:10:08Z", "updated_at": "2026-01-12T07:10:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07263v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2601.07263", "identifiers": {"arxiv": "2601.07263"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.07263"}], "name": "When Bots Take the Bait: Exposing and Mitigating the Emerging Social Engineering Attack in Web Automation Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies deceptive social-engineering web patterns targeting browser automation agents and proposes defenses against bait-based manipulation.", "topics": ["safety-security", "web-navigation", "robustness"], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2601.07263"}
{"aliases": [], "authors": ["Farzad Shami", "Subhrasankha Dey", "Nico Van de Weghe", "Henrikki Tenkanen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68ed822691ecce2f44155653dd7ba80d2ab9e389ed5178bf6e1bc74f69edfc21", "submitted_at": "2026-01-12T09:56:47Z", "updated_at": "2026-01-12T09:56:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07375v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.07375", "identifiers": {"arxiv": "2601.07375"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.07375"}], "name": "GROKE: Vision-Free Navigation Instruction Evaluation via Graph Reasoning on OpenStreetMap", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2601.07375"}
{"aliases": [], "authors": ["Bowen Yang", "Kaiming Jin", "Zhenyu Wu", "Zhaoyang Liu", "Qiushi Sun", "Zehao Li", "JingJing Xie", "Zhoumianze Liu", "Fangzhi Xu", "Kanzhi Cheng", "Qingyun Li", "Yian Wang", "Yu Qiao", "Zun Wang", "Zichen Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cfdaabaf3c6df80eb718e51dc666c54fb2ecf38ec520f6a71ccf5e5ce54248c8", "submitted_at": "2026-01-12T17:55:51Z", "updated_at": "2026-01-12T17:55:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07779v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.07779", "identifiers": {"arxiv": "2601.07779"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.07779"}], "name": "OS-Symphony: A Holistic Framework for Robust and Generalist Computer-Using Agent", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "OS-Symphony coordinates milestone-based reflection and memory with tools that build visually aligned tutorials through browser interaction. The framework evaluates trajectory correction and adaptation on online computer-use benchmarks including OSWorld.", "topics": ["memory", "self-reflection", "planning", "tool-integration"], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2601.07779"}
{"aliases": [], "authors": ["Aarya Doshi", "Yining Hong", "Congying Xu", "Eunsuk Kang", "Alexandros Kapravelos", "Christian Kästner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8aeeca8b8dea9a92ae5ded24db4d4572eaa24ce555da4816e7c30d5477556003", "submitted_at": "2026-01-12T21:31:38Z", "updated_at": "2026-01-12T21:31:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.08012v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.08012", "identifiers": {"arxiv": "2601.08012"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.08012"}], "name": "Towards Verifiably Safe Tool Use for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2601.08012"}
{"aliases": [], "authors": ["Hongjin Qian", "Zhao Cao", "Zheng Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd79963737ff35b5271b7b93127f1b4de403cefb71c7694364e9f662ee567dba", "submitted_at": "2026-01-12T23:44:59Z", "updated_at": "2026-01-12T23:44:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.08079v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.08079", "identifiers": {"arxiv": "2601.08079"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.08079"}], "name": "MemoBrain: Executive Memory as an Agentic Brain for Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2601.08079"}
{"aliases": [], "authors": ["Jungho Cho", "Minbyul Jeong", "Sungrae Park"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ac1ce7d33811b16c2fb630746559be0c9e6667e1ae997ff9f71e801326ef7a1", "submitted_at": "2026-01-13T05:14:09Z", "updated_at": "2026-01-13T05:14:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.08225v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.08225", "identifiers": {"arxiv": "2601.08225"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.08225"}], "name": "User-Oriented Multi-Turn Dialogue Generation with Tool Use at scale", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-13", "url": "https://arxiv.org/abs/2601.08225"}
{"aliases": [], "authors": ["Haoran Su", "Yandong Sun", "Congjia Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9f06aeec35bd10d267841330dd460eb126403fdcc5d6811e7c5a7970c88cc59", "submitted_at": "2026-01-13T05:47:18Z", "updated_at": "2026-01-13T05:47:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.08237v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.08237", "identifiers": {"arxiv": "2601.08237"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.08237"}], "name": "The End of Reward Engineering: How LLMs Are Redefining Multi-Agent Coordination", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-13", "url": "https://arxiv.org/abs/2601.08237"}
{"aliases": [], "authors": ["Xinyi Wu", "Jiagui Chen", "Geng Hong", "Jiayi Dong", "Xudong Pan", "Jiarun Dai", "Min Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "626702d22786336ff65b6e89d97e2034ff7aa144fab4ecc286d169fca6993410", "submitted_at": "2026-01-13T10:21:28Z", "updated_at": "2026-01-13T10:21:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.08406v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.08406", "identifiers": {"arxiv": "2601.08406"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.08406"}], "name": "WebTrap Park: An Automated Platform for Systematic Security Evaluation of Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebTrap Park, a platform that turns web-agent security risks into executable tasks and assesses concrete interactions with live pages. It compares agent frameworks without requiring modifications to their implementations.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2026-01-13", "url": "https://arxiv.org/abs/2601.08406"}
{"aliases": [], "authors": ["Wenyuan Zhang", "Xinghua Zhang", "Haiyang Yu", "Shuaiyi Nie", "Bingli Wu", "Juwei Yue", "Tingwen Liu", "Yongbin Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73df8c5b417c3387530b7730be3aed8b69374c06fdd3573f56147b19359c6912", "submitted_at": "2026-01-13T14:48:34Z", "updated_at": "2026-04-15T14:59:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.08605v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.08605", "identifiers": {"arxiv": "2601.08605"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.08605"}], "name": "ExpSeek: Self-Triggered Experience Seeking for Web Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "ExpSeek uses step-level entropy signals to decide when a web agent should request experience-based guidance. It supplies guidance tailored to the current step and evaluates the intervention across web-agent benchmarks.", "topics": ["memory", "planning", "web-navigation", "evaluation"], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2601.08605"}
{"aliases": [], "authors": ["Yuanlin Duan", "Yuning Wang", "Wenjie Qiu", "He Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6971fc81cb687e8253661fcf7d96d8d9b330a5ef2eec6dc2d46a3cc58fce13b", "submitted_at": "2026-01-13T17:03:31Z", "updated_at": "2026-01-13T17:03:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.08731v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.08731", "identifiers": {"arxiv": "2601.08731"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.08731"}], "name": "Learning from Demonstrations via Capability-Aware Goal Sampling", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-13", "url": "https://arxiv.org/abs/2601.08731"}
{"aliases": [], "authors": ["Zixia Jia", "Jiaqi Li", "Yipeng Kang", "Yuxuan Wang", "Tong Wu", "Quansen Wang", "Xiaobo Wang", "Shuyi Zhang", "Junzhe Shen", "Qing Li", "Siyuan Qi", "Yitao Liang", "Di He", "Zilong Zheng", "Song-Chun Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "70a1e4170ffe969416ba0895be0b94d399b39466771df48a3755de91bf30d896", "submitted_at": "2026-01-14T03:24:08Z", "updated_at": "2026-01-14T03:24:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.09113v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.09113", "identifiers": {"arxiv": "2601.09113"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.09113"}], "name": "The AI Hippocampus: How Far are We From Human Memory?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-14", "url": "https://arxiv.org/abs/2601.09113"}
{"aliases": [], "authors": ["Yibo Lyu", "Gongwei Chen", "Rui Shao", "Weili Guan", "Liqiang Nie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8b697b91942291403491581a61c135c6bfc8b0f3c2a9b9bab0f926af7208fce", "submitted_at": "2026-01-14T17:12:48Z", "updated_at": "2026-05-13T02:54:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.09636v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.09636", "identifiers": {"arxiv": "2601.09636"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.09636"}], "name": "PersonalAlign: Hierarchical Implicit Intent Alignment for Personalized GUI Agent with Long-Term User-Centric Records", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AndroidIntent to evaluate how GUI agents infer omitted preferences and recurring routines from long-term user records. Its HIM-Agent maintains hierarchical memory of preferences and routines for personalized execution and proactive suggestions.", "topics": ["memory", "evaluation", "mobile-navigation"], "updated_at": "2026-05-13", "url": "https://arxiv.org/abs/2601.09636"}
{"aliases": [], "authors": ["Suriya Sureshkumar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b454aeca8e883367c85b3a52196d804d8faaab12ef72f243c20d5e63324acc11", "submitted_at": "2026-01-12T12:13:14Z", "updated_at": "2026-01-12T12:13:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.09749v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.09749", "identifiers": {"arxiv": "2601.09749"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.09749"}], "name": "R-LAM: Reproducibility-Constrained Large Action Models for Scientific Workflow Automation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2601.09749"}
{"aliases": [], "authors": ["Chen Chen", "Jiawei Shao", "Dakuan Lu", "Haoyi Hu", "Xiangcheng Liu", "Hantao Yao", "Wu Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d5ed7e72534fc55d028d4f25cb98b2373c88b622cf76396933d7a49d8ecf2e8", "submitted_at": "2026-01-14T14:27:28Z", "updated_at": "2026-01-14T14:27:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.09770v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.09770", "identifiers": {"arxiv": "2601.09770"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.09770"}], "name": "GUI-Eyes: Tool-Augmented Perception for Visual Grounding in GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Eyes trains GUI grounders to decide when and how to use visual tools such as crops and zooms. Its coarse-to-fine policy uses dense spatial rewards based on location proximity and region overlap, with evaluation on ScreenSpot-Pro.", "topics": ["grounding", "reinforcement-learning", "tool-integration", "screen-understanding"], "updated_at": "2026-01-14", "url": "https://arxiv.org/abs/2601.09770"}
{"aliases": [], "authors": ["Xinxing Ren", "Quagmire Zang", "Caelum Forder", "Suman Deb", "Ahsen Tahir", "Roman J. Georgio", "Peter Carroll", "Zekun Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd53fc2b870e49e9bb99ac65b46ed2ce11e51396b100c82e4723a02e0d1b6eab", "submitted_at": "2026-01-14T21:35:51Z", "updated_at": "2026-01-14T21:35:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.09883v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.09883", "identifiers": {"arxiv": "2601.09883"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.09883"}], "name": "Beyond Rule-Based Workflows: An Information-Flow-Orchestrated Multi-Agents Paradigm via Agent-to-Agent Communication from CORAL", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-14", "url": "https://arxiv.org/abs/2601.09883"}
{"aliases": [], "authors": ["Hanna Foerster", "Tom Blanchard", "Kristina Nikolić", "Ilia Shumailov", "Cheng Zhang", "Robert Mullins", "Nicolas Papernot", "Florian Tramèr", "Yiren Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e32917bc92b65701eea06c09a2c681c7504418f8ce888af2b72012a1da50c9f6", "submitted_at": "2026-01-14T23:06:35Z", "updated_at": "2026-06-04T15:38:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.09923v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.09923", "identifiers": {"arxiv": "2601.09923"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.09923"}], "name": "CaMeLs Can Use Computers Too: System-level Security for Computer Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "NOVA separates trusted task planning from screen-derived observations by producing a branching plan before execution and resolving runtime values through perception. Its OSWorld study examines security and utility while identifying remaining risks from perception-driven branch steering.", "topics": ["safety-security", "planning", "robustness"], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2601.09923"}
{"aliases": [], "authors": ["Saber Zerhoudi", "Michael Granitzer"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf4e40090009768a4cbe46d0701fd1c81513cb14e953476476d1e3180d299697", "submitted_at": "2026-01-14T23:18:54Z", "updated_at": "2026-01-14T23:18:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.09928v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.09928", "identifiers": {"arxiv": "2601.09928"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.09928"}], "name": "In-Browser Agents for Search Assistance", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-14", "url": "https://arxiv.org/abs/2601.09928"}
{"aliases": [], "authors": ["Dian Jiao", "Jiaxin Duan", "Shuai Zhao", "Jiabing Leng", "Yiran Zhang", "Feng Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c9b53f5ff5d058eea463550ad404d1e16cdf60d55473bc032a5143ebd5e29ab", "submitted_at": "2026-01-15T13:29:16Z", "updated_at": "2026-01-17T02:11:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.10378v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.10378", "identifiers": {"arxiv": "2601.10378"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.10378"}], "name": "Global Context Compression with Interleaved Vision-Text Transformation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-17", "url": "https://arxiv.org/abs/2601.10378"}
{"aliases": [], "authors": ["Nadya Abaev", "Denis Klimov", "Gerard Levinov", "David Mimran", "Yuval Elovici", "Asaf Shabtai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14f8cec37db7e4989500e8127e22427bee163d4898bb066073fe1d2f9833c1fe", "submitted_at": "2026-01-15T14:33:36Z", "updated_at": "2026-01-15T14:33:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.10440v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.10440", "identifiers": {"arxiv": "2601.10440"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.10440"}], "name": "AgentGuardian: Learning Access Control Policies to Govern AI Agent Behavior", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-15", "url": "https://arxiv.org/abs/2601.10440"}
{"aliases": [], "authors": ["Fengchao Chen", "Tingmin Wu", "Van Nguyen", "Surya Nepal", "Carsten Rudolph"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9443151152dc7caa3a3789f505f187d2d32aa7475b4b186b84bcc7afba537532", "submitted_at": "2026-01-14T03:29:13Z", "updated_at": "2026-10-10T08:59:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2601.10758v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2601.10758", "identifiers": {"arxiv": "2601.10758"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.10758"}], "name": "When External Threats Become User Context: Tricking Agents into Accepting Malicious Information", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-10", "url": "https://arxiv.org/abs/2601.10758"}
{"aliases": [], "authors": ["Keyu Li", "Junhao Shi", "Yang Xiao", "Mohan Jiang", "Jie Sun", "Yunze Wu", "Dayuan Fu", "Shijie Xia", "Xiaojie Cai", "Tianze Xu", "Weiye Si", "Wenjie Li", "Dequan Wang", "Pengfei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b7c2248c16b00c80eae6376afb924b8e58fe9484cf15ca02efb8c52315ed7ab", "submitted_at": "2026-01-16T07:22:20Z", "updated_at": "2026-04-23T05:01:02Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2601.11044v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2601.11044", "identifiers": {"arxiv": "2601.11044"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.11044"}], "name": "AgencyBench: Benchmarking the Frontiers of Autonomous Agents in 1M-Token Real-World Contexts", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-23", "url": "https://arxiv.org/abs/2601.11044"}
{"aliases": [], "authors": ["Yurun Song", "Jiong Yin", "Rongjunchen Zhang", "Ian G. Harris"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "600f2a44695886fc527b04e1a8b1ad07cec90cf5c58dee834605d5133a760b1e", "submitted_at": "2026-01-14T03:52:57Z", "updated_at": "2026-01-14T03:52:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.11631v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.11631", "identifiers": {"arxiv": "2601.11631"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.11631"}], "name": "Compress to Focus: Efficient Coordinate Compression for Policy Optimization in Multi-Turn GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "CCPO couples GUI policy optimization with coordinate-guided compression of past visual observations. It aggregates targets across rollouts to focus retained context and uses distance-based advantages to train grounding and compression together.", "topics": ["grounding", "reinforcement-learning", "efficiency", "interaction-history"], "updated_at": "2026-01-14", "url": "https://arxiv.org/abs/2601.11631"}
{"aliases": [], "authors": ["Fouad Bousetouane"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "067be1012047e3818a9a5299607e755ebcfa4a1408909f5bdd44e1c5a632d6f3", "submitted_at": "2026-01-15T18:01:59Z", "updated_at": "2026-01-15T18:01:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.11653v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.11653", "identifiers": {"arxiv": "2601.11653"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.11653"}], "name": "AI Agents Need Memory Control Over More Context", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-15", "url": "https://arxiv.org/abs/2601.11653"}
{"aliases": [], "authors": ["Mike A. Merrill", "Alexander G. Shaw", "Nicholas Carlini", "Boxuan Li", "Harsh Raj", "Ivan Bercovich", "Lin Shi", "Jeong Yeon Shin", "Thomas Walshe", "E. Kelly Buchanan", "Junhong Shen", "Guanghao Ye", "Haowei Lin", "Jason Poulos", "Maoyu Wang", "Marianna Nezhurina", "Jenia Jitsev", "Di Lu", "Orfeas Menis Mastromichalakis", "Zhiwei Xu", "Zizhao Chen", "Yue Liu", "Robert Zhang", "Leon Liangyu Chen", "Anurag Kashyap", "Jan-Lucas Uslu", "Jeffrey Li", "Jianbo Wu", "Minghao Yan", "Song Bian", "Vedang Sharma", "Ke Sun", "Steven Dillmann", "Akshay Anand", "Andrew Lanpouthakoun", "Bardia Koopah", "Changran Hu", "Etash Guha", "Gabriel H. S. Dreiman", "Jiacheng Zhu", "Karl Krauth", "Li Zhong", "Niklas Muennighoff", "Robert Amanfu", "Shangyin Tan", "Shreyas Pimpalgaonkar", "Tushar Aggarwal", "Xiangning Lin", "Xin Lan", "Xuandong Zhao", "Yiqing Liang", "Yuanli Wang", "Zilong Wang", "Changzhi Zhou", "David Heineman", "Hange Liu", "Harsh Trivedi", "John Yang", "Junhong Lin", "Manish Shetty", "Michael Yang", "Nabil Omi", "Negin Raoof", "Shanda Li", "Terry Yue Zhuo", "Wuwei Lin", "Yiwei Dai", "Yuxin Wang", "Wenhao Chai", "Shang Zhou", "Dariush Wahdany", "Ziyu She", "Jiaming Hu", "Zhikang Dong", "Yuxuan Zhu", "Sasha Cui", "Ahson Saiyed", "Arinbjörn Kolbeinsson", "Jesse Hu", "Christopher Michael Rytting", "Ryan Marten", "Yixin Wang", "Alex Dimakis", "Andy Konwinski", "Ludwig Schmidt"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40586833a0a14c29a39920f0e44e8d375191af28fa96c2f6a29aabe1ed7d3c45", "submitted_at": "2026-01-17T01:29:30Z", "updated_at": "2026-01-17T01:29:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.11868v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "id": "arxiv:2601.11868", "identifiers": {"arxiv": "2601.11868"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.11868"}], "name": "Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-17", "url": "https://arxiv.org/abs/2601.11868"}
{"aliases": [], "authors": ["Adarsh Kumarappan", "Pareesa Ameneh Golnari", "Wen Wen", "Xiaoyu Liu", "Gabriel Ryan", "Yuting Sun", "Shengyu Fu", "Elsie Nallipogu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "59d2b1f803636d4e75364710638704e909e414c6249cef26d42a4ff5a5da305b", "submitted_at": "2026-01-17T03:33:08Z", "updated_at": "2026-05-16T19:19:54Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.11895v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.11895", "identifiers": {"arxiv": "2601.11895"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.11895"}], "name": "DevBench: A Realistic, Developer-Informed Benchmark for Code Generation Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-16", "url": "https://arxiv.org/abs/2601.11895"}
{"aliases": [], "authors": ["YenTing Lee", "Keerthi Koneru", "Zahra Moslemi", "Sheethal Kumar", "Ramesh Radhakrishnan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e366da130437eb52747be78192db9a7d05800ff453a7af830e7f53cadf436dd", "submitted_at": "2026-01-17T04:09:02Z", "updated_at": "2026-01-17T04:09:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.11903v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.11903", "identifiers": {"arxiv": "2601.11903"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.11903"}], "name": "AEMA: Verifiable Evaluation Framework for Trustworthy and Controlled Agentic LLM Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-17", "url": "https://arxiv.org/abs/2601.11903"}
{"aliases": [], "authors": ["Xinmeng Hou", "Peiliang Gong", "Bohao Qu", "Wuqi Wang", "Qing Guo", "Yang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f64728faa86136d794b34c589647fa4d2158de3279d41a38755726cf3c36bc73", "submitted_at": "2026-01-17T09:12:26Z", "updated_at": "2026-01-17T09:12:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.11974v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.11974", "identifiers": {"arxiv": "2601.11974"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.11974"}], "name": "Learn Like Humans: Use Meta-cognitive Reflection for Efficient Self-Improvement", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-17", "url": "https://arxiv.org/abs/2601.11974"}
{"aliases": [], "authors": ["Xiaomei Zhang", "Zhaoxi Zhang", "Leo Yu Zhang", "Yanjun Zhang", "Guanhong Tao", "Shirui Pan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "437c6784e206b28e0c5312ec026298acf5113381b57c2f751072abc63741ec42", "submitted_at": "2026-01-17T13:02:41Z", "updated_at": "2026-01-17T13:02:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.12042v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.12042", "identifiers": {"arxiv": "2601.12042"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.12042"}], "name": "Less Is More -- Until It Breaks: Security Pitfalls of Vision Token Compression in Large Vision-Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-17", "url": "https://arxiv.org/abs/2601.12042"}
{"aliases": [], "authors": ["Jiawei Xu", "Arief Koesdwiady", "Sisong Bei", "Yan Han", "Baixiang Huang", "Dakuo Wang", "Yutong Chen", "Zheshen Wang", "Peihao Wang", "Pan Li", "Ying Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bc55ee0d4ee1e2e28d3ff03e42cbaca423b6a7743a4dae6faf9cf6317f358ae0", "submitted_at": "2026-01-18T08:16:09Z", "updated_at": "2026-01-18T08:16:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.12307v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.12307", "identifiers": {"arxiv": "2601.12307"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.12307"}], "name": "Rethinking the Value of Multi-Agent Workflow: A Strong Single Agent Baseline", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-18", "url": "https://arxiv.org/abs/2601.12307"}
{"aliases": [], "authors": ["Yi Qian", "Kunwei Qian", "Xingbang He", "Ligeng Chen", "Jikang Zhang", "Tiantai Zhang", "Haiyang Wei", "Linzhang Wang", "Hao Wu", "Bing Mao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "21848dea9e000125550fff7fe3844a57bf100215bf9eb800cbdbb612675f7074", "submitted_at": "2026-01-18T10:54:54Z", "updated_at": "2026-07-15T15:32:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.12349v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.12349", "identifiers": {"arxiv": "2601.12349"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.12349"}], "name": "Mind the Gap: Action Rebinding Attacks against Android GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies action-rebinding attacks in which an Android GUI agent’s planned input is redirected to a different application during the observation–action delay. The evaluation examines cross-application execution and the consequences of treating an earlier screen as the current action target.", "topics": ["safety-security", "cross-app", "evaluation"], "updated_at": "2026-07-15", "url": "https://arxiv.org/abs/2601.12349"}
{"aliases": [], "authors": ["Roy Betser", "Amit Giloni", "Shamik Bose", "Sindhu Padakandla", "Chiara Picardi", "Lidor Erez", "Roman Vainshtein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4c876f51d4450e5fae0a3b4df1e502ee5d1918e0c0b542b6112226986e44480", "submitted_at": "2026-01-18T15:10:18Z", "updated_at": "2026-08-30T13:37:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.12449v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2601.12449", "identifiers": {"arxiv": "2601.12449"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.12449"}], "name": "AgenTRIM: Tool Risk Mitigation for Agentic AI", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-30", "url": "https://arxiv.org/abs/2601.12449"}
{"aliases": [], "authors": ["Arunkumar V", "Gangadharan G. R.", "Rajkumar Buyya"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3de4529325c90a0addd06938f81876283dc9a776e12ee7038efea5dda2a91e9b", "submitted_at": "2026-01-18T19:51:16Z", "updated_at": "2026-01-18T19:51:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.12560v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.12560", "identifiers": {"arxiv": "2601.12560"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.12560"}], "name": "Agentic Artificial Intelligence (AI): Architectures, Taxonomies, and Evaluation of Large Language Model Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-18", "url": "https://arxiv.org/abs/2601.12560"}
{"aliases": [], "authors": ["Xiangyu Shi", "Junyang Ding", "Xu Zhao", "Sinong Zhan", "Payal Mohapatra", "Daniel Quispe", "Kojo Welbeck", "Jian Cao", "Wei Chen", "Ping Guo", "Qi Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cce08a032b452aa30b0d71a171612ad3cde2a57c344953e2fb2c8f69694c2f09", "submitted_at": "2026-01-19T01:10:49Z", "updated_at": "2026-09-27T07:54:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.12641v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.12641", "identifiers": {"arxiv": "2601.12641"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.12641"}], "name": "STEP-LLM: Generating CAD STEP Models from Natural Language with Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-27", "url": "https://arxiv.org/abs/2601.12641"}
{"aliases": [], "authors": ["Wenqi Zhang", "Yulin Shen", "Changyue Jiang", "Jiarun Dai", "Geng Hong", "Xudong Pan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b57586682d4675387dea2687e7c15d70cc10aecac883e35fe1812c39abcbc1dc", "submitted_at": "2026-01-19T08:32:09Z", "updated_at": "2026-01-19T08:32:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.12822v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.12822", "identifiers": {"arxiv": "2601.12822"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.12822"}], "name": "MirrorGuard: Toward Secure Computer-Use Agents via Simulation-to-Real Reasoning Correction", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "MirrorGuard trains a reasoning-correction defense using GUI risk trajectories generated in a text-based neural-symbolic simulator. It evaluates whether corrections transfer to computer-use agents operating real interfaces, keeping simulated training separate from deployment tests.", "topics": ["safety-security", "data-generation", "robustness"], "updated_at": "2026-01-19", "url": "https://arxiv.org/abs/2601.12822"}
{"aliases": [], "authors": ["Zecheng Li", "Zhihui Cao", "Wenke Huang", "Yudong Zhang", "Keying Qi", "Rui Wang", "Zeyu Zheng", "Jian Zhao", "Hao Zhu", "Hengxin Wu", "Yuran Wang", "Guitao Fan", "Guokun Wu", "Yicong Liu", "Zhilin Gao", "Haikun Xu", "He Yang", "Minqi Xiang", "Xingyu Liu", "Zuojian Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa8cb4137e862bf8031a55bc8f81e1d6c903733f2cfac9cf758e5f2f92884481", "submitted_at": "2026-01-19T13:50:43Z", "updated_at": "2026-01-19T13:50:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.13060v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.13060", "identifiers": {"arxiv": "2601.13060"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.13060"}], "name": "MagicGUI-RMS: A Multi-Agent Reward Model System for Self-Evolving GUI Agents via Automated Feedback Reflux", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "MagicGUI-RMS combines domain-specific and general reward models to assess GUI actions and propose corrections. A structured data pipeline and feedback-recycling process support repeated reward-model and agent improvement.", "topics": ["reward-models", "data-generation", "evaluation", "robustness"], "updated_at": "2026-01-19", "url": "https://arxiv.org/abs/2601.13060"}
{"aliases": [], "authors": ["Diego Gosmar", "Deborah A. Dahl"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbcba59f27d78c1688563e0bc35e5f75c34e2810ec1757a4b062e6de2edc8fd2", "submitted_at": "2026-01-19T16:10:11Z", "updated_at": "2026-01-19T16:10:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.13186v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.13186", "identifiers": {"arxiv": "2601.13186"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.13186"}], "name": "Prompt Injection Mitigation with Agentic AI, Nested Learning, and AI Sustainability via Semantic Caching", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-19", "url": "https://arxiv.org/abs/2601.13186"}
{"aliases": [], "authors": ["Akbar Anbar Jafari", "Cagri Ozcinar", "Gholamreza Anbarjafari"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "560bed004df8dee0c579eceb754b6b9ce889679e47efca51f84e300bb47229e8", "submitted_at": "2026-01-19T20:33:26Z", "updated_at": "2026-01-19T20:33:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.13383v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "id": "arxiv:2601.13383", "identifiers": {"arxiv": "2601.13383"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.13383"}], "name": "A Lightweight Modular Framework for Constructing Autonomous Agents Driven by Large Language Models: Design, Implementation, and Applications in AgentForge", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-19", "url": "https://arxiv.org/abs/2601.13383"}
{"aliases": [], "authors": ["Amine Rostane"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "180fe6a86a5f0f6ad94e70e9a4575de10ea8c5687b04083aee42d3b8ef1f4865", "submitted_at": "2026-01-19T23:37:10Z", "updated_at": "2026-01-19T23:37:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.13462v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.13462", "identifiers": {"arxiv": "2601.13462"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.13462"}], "name": "SpatialBench-UC: Uncertainty-Aware Evaluation of Spatial Prompt Following in Text-to-Image Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-19", "url": "https://arxiv.org/abs/2601.13462"}
{"aliases": [], "authors": ["Egor Cherepanov", "Daniil Zelezetsky", "Alexey K. Kovalev", "Aleksandr I. Panov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b324c47a674891e311ff2e0a50ad6e1a7188c2c43cb5ffe52bee6e8fc82b023", "submitted_at": "2026-01-20T18:44:28Z", "updated_at": "2026-07-01T10:11:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.14232v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.14232", "identifiers": {"arxiv": "2601.14232"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.14232"}], "name": "KAGE-Bench: Fast Known-Axis Visual Generalization Evaluation for Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-01", "url": "https://arxiv.org/abs/2601.14232"}
{"aliases": [], "authors": ["Rishit Chugh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45585e2e30e60c3f62d55ce157f9ec2a6935d63880181cd28274f9f98f08efbf", "submitted_at": "2026-01-20T06:01:02Z", "updated_at": "2026-01-20T06:01:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.15331v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.15331", "identifiers": {"arxiv": "2601.15331"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.15331"}], "name": "RECAP: A Resource-Efficient Method for Adversarial Prompting in Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-20", "url": "https://arxiv.org/abs/2601.15331"}
{"aliases": [], "authors": ["Taofeng Xue", "Chong Peng", "Mianqiu Huang", "Linsen Guo", "Tiancheng Han", "Haozhe Wang", "Jianing Wang", "Xiaocheng Zhang", "Xin Yang", "Dengchang Zhao", "Jinrui Ding", "Xiandi Ma", "Yuchen Xie", "Peng Pei", "Xunliang Cai", "Xipeng Qiu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "38b2798ce9ebd69a46202f0fa8ceca859394acda4f5173b5e30248914c8c35ac", "submitted_at": "2026-01-22T11:36:43Z", "updated_at": "2026-01-23T08:52:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.15876v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.15876", "identifiers": {"arxiv": "2601.15876"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.15876"}], "name": "EvoCUA: Evolving Computer Use Agents via Learning from Scalable Synthetic Experience", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "EvoCUA combines task synthesis with executable validators, asynchronous sandbox rollouts and iterative policy updates. It turns successful routines and analyzed failures into supervision and evaluates the evolving agent on OSWorld.", "topics": ["data-generation", "sandbox", "long-horizon", "self-reflection"], "updated_at": "2026-01-23", "url": "https://arxiv.org/abs/2601.15876"}
{"aliases": [], "authors": ["Daixuan Cheng", "Shaohan Huang", "Yuxian Gu", "Huatong Song", "Guoxin Chen", "Li Dong", "Wayne Xin Zhao", "Ji-Rong Wen", "Furu Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6418171ce26fffaa9debd94806b28d6018e5485ff991a359485558cebfefaebb", "submitted_at": "2026-01-22T18:57:09Z", "updated_at": "2026-04-08T06:37:40Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.16206v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2601.16206", "identifiers": {"arxiv": "2601.16206"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.16206"}], "name": "Computer Environments Elicit General Agentic Intelligence in LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-08", "url": "https://arxiv.org/abs/2601.16206"}
{"aliases": [], "authors": ["Simret Araya Gebreegziabher", "Yukun Yang", "Charles Chiang", "Hojun Yoo", "Chaoran Chen", "Hyo Jin Do", "Zahra Ashktorab", "Werner Geyer", "Diego Gómez-Zará", "Toby Jia-Jun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bdc29648446cd79e0c0b9f72814457b55258d501283ce2fa2dedcd8262f885b", "submitted_at": "2026-01-22T22:42:40Z", "updated_at": "2026-01-22T22:42:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.16356v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.16356", "identifiers": {"arxiv": "2601.16356"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.16356"}], "name": "The Behavioral Fabric of LLM-Powered GUI Agents: Human Values and Interaction Outcomes", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a controlled web-task testbed for studying how stated preferences and human values influence GUI-agent decisions. It analyzes how personas alter navigation and how prominent interface cues can override the intended preference guidance.", "topics": ["evaluation", "ui-design", "human-in-the-loop"], "updated_at": "2026-01-22", "url": "https://arxiv.org/abs/2601.16356"}
{"aliases": [], "authors": ["Mingxian Yu", "Siqi Luo", "Xu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ce0909d1c9081f099ae0dc0b0182e0442fd71fd58c62b13d486d154afc2187e", "submitted_at": "2026-01-24T11:35:37Z", "updated_at": "2026-01-24T11:35:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.17418v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2601.17418", "identifiers": {"arxiv": "2601.17418"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.17418"}], "name": "GraphPilot: GUI Task Automation with One-Step LLM Reasoning Powered by Knowledge Graph", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "GraphPilot builds app-specific knowledge graphs from explored pages, elements and transitions, then uses them to plan action sequences. It requests additional interface observations when uncertain and checks plans against graph transition rules before correction and execution.", "topics": ["mobile-navigation", "planning", "memory", "efficiency"], "updated_at": "2026-01-24", "url": "https://arxiv.org/abs/2601.17418"}
{"aliases": [], "authors": ["Narek Maloyan", "Dmitry Namiot"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8d1bc1c63517b92ca15ba74fe51546a968f46e49bab3370b011464ea7322622f", "submitted_at": "2026-01-24T18:39:21Z", "updated_at": "2026-01-24T18:39:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.17548v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.17548", "identifiers": {"arxiv": "2601.17548"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.17548"}], "name": "Prompt Injection Attacks on Agentic Coding Assistants: A Systematic Analysis of Vulnerabilities in Skills, Tools, and Protocol Ecosystems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-24", "url": "https://arxiv.org/abs/2601.17548"}
{"aliases": [], "authors": ["Ying Mo", "Yu Bai", "Dapeng Sun", "Yuqian Shi", "Yukai Miao", "Li Chen", "Dan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa7c8ce1b0e584c5279691ff06cd522d7c8ea1f4801d7074324c6845ce482efb", "submitted_at": "2026-01-25T06:58:15Z", "updated_at": "2026-01-25T06:58:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.17722v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.17722", "identifiers": {"arxiv": "2601.17722"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.17722"}], "name": "EntWorld: A Holistic Environment and Benchmark for Verifiable Enterprise GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces EntWorld, an enterprise GUI benchmark built by deriving task logic from database schemas. Long workflows across enterprise domains are assessed with deterministic SQL-based state checks rather than screenshot similarity alone.", "topics": ["evaluation", "industry", "workflow"], "updated_at": "2026-01-25", "url": "https://arxiv.org/abs/2601.17722"}
{"aliases": [], "authors": ["Nan Chen", "Jing Lu", "Zilong Wang", "Luna K. Qiu", "Siming Chen", "Yuqing Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae1716f673e47cd0e61b5f2225ca4871dfd8522e35037fc8d226f9c7cd0e3e75", "submitted_at": "2026-01-26T03:02:26Z", "updated_at": "2026-01-26T03:02:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.18092v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.18092", "identifiers": {"arxiv": "2601.18092"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.18092"}], "name": "From Struggle to Success: Context-Aware Guidance for Screen Reader Users in Computer Use", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-26", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces AskEase, an on-demand assistant that combines contextual information to provide screen-reader-friendly computer-use guidance. A study with screen-reader users evaluates task completion and perceived workload while preserving human control of the interaction.", "topics": ["human-in-the-loop", "evaluation", "workflow"], "updated_at": "2026-01-26", "url": "https://arxiv.org/abs/2601.18092"}
{"aliases": [], "authors": ["Dezhang Kong", "Zhuxi Wu", "Shiqi Liu", "Zhicheng Tan", "Kuichen Lu", "Minghao Li", "Qichen Liu", "Shengyu Chu", "Zhenhua Xu", "Xuan Liu", "Meng Han"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ae4de775fad8e66055667175467487cb035da829f5e867a9ddb9d90dfe275e1", "submitted_at": "2026-01-26T03:58:10Z", "updated_at": "2026-03-13T04:12:36Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.18113v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2601.18113", "identifiers": {"arxiv": "2601.18113"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.18113"}], "name": "MalURLBench: A Benchmark Evaluating Agents' Vulnerabilities When Processing Web URLs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-13", "url": "https://arxiv.org/abs/2601.18113"}
{"aliases": [], "authors": ["Shaokang Wang", "Pei Fu", "Ruoceng Zhang", "Shaojie Zhang", "Xiuwen Xi", "Jiahui Yang", "Bin Qin", "Ying Huang", "Zhenbo Luo", "Jian Luan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccb46b86b0c0b854ed2f1e17f9e12bf031c5ed1349f3d3c44d3276b2a3a195ee", "submitted_at": "2026-01-26T06:29:41Z", "updated_at": "2026-06-26T07:19:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.18197v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2601.18197", "identifiers": {"arxiv": "2601.18197"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.18197"}], "name": "GAIA: A Data Flywheel System for Training GUI Test-Time Scaling Critic Models", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "GAIA trains a GUI action critic on positive and negative examples, then uses that critic to collect improved examples for another training round. It studies how recycling these data affects the test-time action choices of different GUI agents.", "topics": ["reward-models", "data-generation", "evaluation"], "updated_at": "2026-06-26", "url": "https://arxiv.org/abs/2601.18197"}
{"aliases": [], "authors": ["Xuan Wang", "Siyuan Su", "Quantong Fu", "Yongxiang Hu", "Yangfan Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cb90eed33e989d38cd4058d54ba7d073f1b8e5c1453a9106dfe5e533da7b08a1", "submitted_at": "2026-01-26T09:35:10Z", "updated_at": "2026-08-20T07:50:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.18305v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.18305", "identifiers": {"arxiv": "2601.18305"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.18305"}], "name": "SwipeGen: Bridging the Execution Gap in GUI Agents via Human-like Swipe Synthesis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SwipeGen for synthesizing varied swipe trajectories and SwipeBench for assessing their execution quality. The study examines how richer control of trajectory, speed, and timing changes GUI agents’ interaction performance.", "topics": ["action-prediction", "data-generation", "evaluation"], "updated_at": "2026-08-20", "url": "https://arxiv.org/abs/2601.18305"}
{"aliases": [], "authors": ["Ji Zeng", "Dayuan Fu", "Tiantian Mi", "Yumin Zhuang", "Yaxing Huang", "Xuefeng Li", "Lyumanshan Ye", "Muhang Xie", "Qishuo Hua", "Zhen Huang", "Mohan Jiang", "Hanning Wang", "Jifan Lin", "Yang Xiao", "Jie Sun", "Yunze Wu", "Pengfei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "113546b6f84274fb5c1db6b81cb86d39331305a869a2237e5d5d25fb1c8161b2", "submitted_at": "2026-01-26T12:20:18Z", "updated_at": "2026-01-27T12:16:16Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.18418v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.18418", "identifiers": {"arxiv": "2601.18418"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.18418"}], "name": "daVinci-Dev: Agent-native Mid-training for Software Engineering", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-27", "url": "https://arxiv.org/abs/2601.18418"}
{"aliases": [], "authors": ["Dongrui Liu", "Qihan Ren", "Chen Qian", "Shuai Shao", "Yuejin Xie", "Yu Li", "Zhonghao Yang", "Haoyu Luo", "Peng Wang", "Qingyu Liu", "Binxin Hu", "Ling Tang", "Jilin Mei", "Dadi Guo", "Leitao Yuan", "Junyao Yang", "Guanxu Chen", "Qihao Lin", "Yi Yu", "Bo Zhang", "Jiaxuan Guo", "Jie Zhang", "Wenqi Shao", "Huiqi Deng", "Zhiheng Xi", "Wenjie Wang", "Wenxuan Wang", "Wen Shen", "Zhikai Chen", "Haoyu Xie", "Jialing Tao", "Juntao Dai", "Jiaming Ji", "Zhongjie Ba", "Linfeng Zhang", "Yong Liu", "Quanshi Zhang", "Lei Zhu", "Zhihua Wei", "Hui Xue", "Chaochao Lu", "Jing Shao", "Xia Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "be1b49dd9b603104cac9c3374463812d099245c0f53bd45b4af2d14daaa57b99", "submitted_at": "2026-01-26T13:45:41Z", "updated_at": "2026-04-23T11:55:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.18491v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.18491", "identifiers": {"arxiv": "2601.18491"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.18491"}], "name": "AgentDoG: A Diagnostic Guardrail Framework for AI Agent Safety and Security", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-23", "url": "https://arxiv.org/abs/2601.18491"}
{"aliases": [], "authors": ["Jens Kohl", "Otto Kruse", "Youssef Mostafa", "Andre Luckow", "Karsten Schroer", "Thomas Riedl", "Ryan French", "David Katz", "Manuel P. Luitz", "Tanrajbir Takher", "Ken E. Friedl", "Céline Laurent-Winter"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a38f8b41bf75138f95973d9d650ee83c82e651abd6a70f05d3f08e2d34eb3c8", "submitted_at": "2026-01-25T11:52:30Z", "updated_at": "2026-01-25T11:52:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.18827v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.18827", "identifiers": {"arxiv": "2601.18827"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.18827"}], "name": "Automated structural testing of LLM-based agents: methods, framework, and case studies", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-25", "url": "https://arxiv.org/abs/2601.18827"}
{"aliases": [], "authors": ["Yanxi Wang", "Zhiling Zhang", "Wenbo Zhou", "Weiming Zhang", "Jie Zhang", "Qiannan Zhu", "Yu Shi", "Shuxin Zheng", "Jiyan He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ca912ad67a681e8e908a5a403e982f39b5080bee4df376eb0b598468f3bb504", "submitted_at": "2026-01-26T11:33:40Z", "updated_at": "2026-05-13T07:11:36Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.18842v3"}}, "environments": ["android", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.18842", "identifiers": {"arxiv": "2601.18842"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.18842"}], "name": "GUIGuard-Bench: Toward a General Evaluation for Privacy-Preserving GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUIGuard-Bench, which annotates private regions in GUI trajectories with categories, risk levels, and relevance to the user’s task. It evaluates privacy recognition and the effect of screenshot protection on offline planning and task utility.", "topics": ["safety-security", "evaluation", "grounding"], "updated_at": "2026-05-13", "url": "https://arxiv.org/abs/2601.18842"}
{"aliases": [], "authors": ["Libo Sun", "Jiwen Zhang", "Siyuan Wang", "Zhongyu Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1930b4efca7ee38fc6cd1ceb283c14ed9b1388e7b758b951ba9469ed3067c3e0", "submitted_at": "2026-01-27T04:58:48Z", "updated_at": "2026-01-31T15:17:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.19199v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.19199", "identifiers": {"arxiv": "2601.19199"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.19199"}], "name": "MAGNET: Towards Adaptive GUI Agents with Memory-Driven Knowledge Evolution", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "MAGNET maintains separate memories for stable interface functions and task procedures across changing layouts and workflows. It updates frequently used knowledge and evaluates adaptation in AndroidWorld and offline distribution-shift settings.", "topics": ["memory", "grounding", "robustness", "evaluation"], "updated_at": "2026-01-31", "url": "https://arxiv.org/abs/2601.19199"}
{"aliases": [], "authors": ["Sijia Li", "Xiaoyu Tan", "Shahir Ali", "Niels Schmidt", "Gengchen Ma", "Xihe Qiu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "223b9a915e5ea9849644318e1e5426e64ad8655ab46d0b007c30e1967130dc50", "submitted_at": "2026-01-27T07:46:05Z", "updated_at": "2026-01-27T07:46:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.19306v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:04.629294+00:00", "role": "primary", "sha256": "53521b0a80976320b10efcf7840629e4a2085119710324dd48db2cfd9c5d9cf6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:04.629294+00:00", "role": "primary", "sha256": "53521b0a80976320b10efcf7840629e4a2085119710324dd48db2cfd9c5d9cf6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:20.444793+00:00", "role": "primary", "sha256": "ef3bbc8adcc147239ea910e0c59904d7a26ee1610a7a4d1613be2cb845098536", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2601.19306v1"}], "id": "arxiv:2601.19306", "identifiers": {"arxiv": "2601.19306"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:04.629294+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.19306"}], "name": "Curiosity Driven Knowledge Retrieval for Mobile Agents", "notes": "The curiosity signal is a model-based proxy, not a calibrated probability of failure. Accurate external app information is a dependency, and AndroidWorld is the reported evaluation environment. No performance numbers were independently reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3.1 Problem Formulation; 3.2 Curiosity-Driven Information Gain Estimation; 3.3 Retrieving and Embedding External Knowledge); not an entire-paper reading or reproduction claim.", "published_at": "2026-01-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses execution uncertainty to trigger retrieval from application documentation, code, and past trajectories, organizing the results into structured AppCards for Android agent planning.", "topics": ["memory", "planning", "mobile-navigation"], "updated_at": "2026-01-27", "url": "https://arxiv.org/abs/2601.19306"}
{"aliases": [], "authors": ["Shuoxin Wang", "Chang Liu", "Gowen Loo", "Lifan Zheng", "Kaiwen Wei", "Xinyi Zeng", "Jingyuan Zhang", "Yu Tian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dbe8e4ef8611f99058d6a23d2c414af318b672e8369bebf47e633830ab0652ff", "submitted_at": "2026-01-28T01:44:19Z", "updated_at": "2026-01-28T01:44:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.20162v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2601.20162", "identifiers": {"arxiv": "2601.20162"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.20162"}], "name": "Me-Agent: A Personalized Mobile Agent with Two-Level User Habit Learning for Enhanced Interaction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-28", "url": "https://arxiv.org/abs/2601.20162"}
{"aliases": [], "authors": ["Kazuma Yamasaki", "Joseph Ayobami Joshua", "Tasha Settewong", "Mahmoud Alfadel", "Kazumasa Shimari", "Kenichi Matsumoto"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c7db1d151d944df8c3b4090af05fd13a1685e9a2d16753afaaedaac683edbbf2", "submitted_at": "2026-01-28T02:11:34Z", "updated_at": "2026-01-28T02:11:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.20171v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.20171", "identifiers": {"arxiv": "2601.20171"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.20171"}], "name": "Who Writes the Docs in SE 3.0? Agent vs. Human Documentation Pull Requests", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-28", "url": "https://arxiv.org/abs/2601.20171"}
{"aliases": [], "authors": ["Qinzhuo Wu", "Zhizhuo Yang", "Hanhao Li", "Pengzhi Gao", "Wei Liu", "Jian Luan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1cbf16dd3c790996758e00c085a9c2bccabf5b24f945bbd8440cc1f953e38568", "submitted_at": "2026-01-28T07:49:48Z", "updated_at": "2026-01-29T02:32:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.20335v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.20335", "identifiers": {"arxiv": "2601.20335"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.20335"}], "name": "MobileBench-OL: A Comprehensive Chinese Benchmark for Evaluating Mobile GUI Agents in Real-World Environment", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MobileBench-OL, an online benchmark of Chinese-app tasks spanning execution, reasoning, exploration, and noisy conditions. Its automatic evaluation and reset framework supports repeated testing of mobile GUI agents in real environments.", "topics": ["evaluation", "mobile-navigation", "robustness"], "updated_at": "2026-01-29", "url": "https://arxiv.org/abs/2601.20335"}
{"aliases": [], "authors": ["Le Zhang", "Yixiong Xiao", "Xinjiang Lu", "Jingjia Cao", "Yusai Zhao", "Jingbo Zhou", "Lang An", "Zikan Feng", "Wanxiang Sha", "Yu Shi", "Congxi Xiao", "Jian Xiong", "Yankai Zhang", "Hua Wu", "Haifeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ad764cf86a2ecab81f3f730d6d26ea582544a355a90a8be55be220c98aabd7bf", "submitted_at": "2026-01-28T08:45:17Z", "updated_at": "2026-01-28T08:45:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.20380v1"}}, "environments": ["android", "desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.20380", "identifiers": {"arxiv": "2601.20380"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.20380"}], "name": "OmegaUse: Building a General-Purpose GUI Agent for Autonomous Task Execution", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "OmegaUse combines exploratory and taxonomy-guided data synthesis with supervised fine-tuning followed by GRPO for GUI control. The report introduces OS-Nav for offline Android and Ubuntu evaluation alongside established grounding and action-prediction benchmarks.", "topics": ["data-generation", "reinforcement-learning", "grounding", "evaluation"], "updated_at": "2026-01-28", "url": "https://arxiv.org/abs/2601.20380"}
{"aliases": [], "authors": ["Kaiyuan Chen", "Qimin Wu", "Taiyu Hou", "Tianhao Tang", "Xueyu Hu", "Yuchen Hou", "Bikun Li", "Chengming Qian", "Guoyin Wang", "Haolin Chen", "Haotong Tian", "Haoye Zhang", "Haoyu Bian", "Hongbing Pan", "Hongkang Zhang", "Hongyi Zhou", "Jiaqi Cai", "Jiewu Rao", "Jiyuan Ren", "Keduan Huang", "Lucia Zhu Huang", "Mingyu Yuan", "Naixu Guo", "Qicheng Tang", "Qinyan Zhang", "Shuai Chen", "Siheng Chen", "Ting Ting Li", "Xiaoxing Guo", "Yaocheng Zuo", "Yaoqi Guo", "Yinan Wang", "Yinzhou Yu", "Yize Wang", "Yuan Jiang", "Yuan Tian", "Yuanshuo Zhang", "Yuxuan Liu", "Yvette Yan Zeng", "Zenyu Shan", "Zihan Yin", "Xiaobo Hu", "Yang Liu", "Yixin Ren", "Yuan Gong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75d0bd608e6108d8d64c1da1aa998661f78997b6483c3cac29c08da286b20cb4", "submitted_at": "2026-01-28T13:49:18Z", "updated_at": "2026-01-30T13:36:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.20613v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.20613", "identifiers": {"arxiv": "2601.20613"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.20613"}], "name": "AgentIF-OneDay: A Task-level Instruction-Following Benchmark for General AI Agents in Daily Scenarios", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-30", "url": "https://arxiv.org/abs/2601.20613"}
{"aliases": [], "authors": ["Jing Wu", "Wenjie Ai", "Daphne Barretto", "Yiye Chen", "Qingyu Chen", "Yuhang He", "Pranit Chawla", "Nicholas Gydé", "Yanan Jian", "Vibhav Vineet"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0139f7a1402d8dbfdecf65cc53c032495b2084f24cc38ff43841642c9102f29c", "submitted_at": "2026-01-28T14:35:23Z", "updated_at": "2026-08-26T22:31:44Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.20650v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.20650", "identifiers": {"arxiv": "2601.20650"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.20650"}], "name": "OS-Marathon: Benchmarking Computer-Use Agents on Vast-Horizon, Repetitive Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces OS-Marathon to evaluate long repetitive computer-use workflows whose execution grows with the amount of input data. It studies failure accumulation under task decomposition and GraphDemo, which adapts agents using a demonstration of the recurring workflow.", "topics": ["evaluation", "long-horizon", "human-in-the-loop"], "updated_at": "2026-08-26", "url": "https://arxiv.org/abs/2601.20650"}
{"aliases": [], "authors": ["Ziwei Liu", "Borui Kang", "Hangjie Yuan", "Zixiang Zhao", "Wei Li", "Yifan Zhu", "Tao Feng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1083cf063ccf1781f2cdfe5e266b75b52079fd12ad16d05755577b880c010aec", "submitted_at": "2026-01-28T16:06:31Z", "updated_at": "2026-03-25T10:25:50Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2601.20732v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.20732", "identifiers": {"arxiv": "2601.20732"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.20732"}], "name": "Continual GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Continual GUI Agents defines grounding under changing application domains and screen resolutions. Its GUI-AiF training framework uses separate point- and region-alignment rewards to adapt grounding as these distributions shift.", "topics": ["grounding", "reinforcement-learning", "robustness", "evaluation"], "updated_at": "2026-03-25", "url": "https://arxiv.org/abs/2601.20732"}
{"aliases": [], "authors": ["Xin Jin", "Jinming Liu", "Yuntao Wei", "Junyan Lin", "Zhicheng Wang", "Jianguo Huang", "Xudong Yang", "Yanxiao Liu", "Wenjun Zeng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51b52155ffb391c524b5263965eef5e6a518a9d74c5899a57237edb529daeeeb", "submitted_at": "2026-01-28T16:18:20Z", "updated_at": "2026-01-28T16:18:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.20742v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.20742", "identifiers": {"arxiv": "2601.20742"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.20742"}], "name": "Compression Tells Intelligence: Visual Coding, Visual Token Technology, and the Unification", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-28", "url": "https://arxiv.org/abs/2601.20742"}
{"aliases": [], "authors": ["Tianyi Chen", "Yinheng Li", "Michael Solodko", "Sen Wang", "Nan Jiang", "Tingyuan Cui", "Junheng Hao", "Jongwoo Ko", "Sara Abdali", "Leon Xu", "Suzhen Zheng", "Hao Fan", "Pashmina Cameron", "Justin Wagle", "Kazuhito Koishida"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a3efa3cd0add437e327ea49896a3ee794daa178664dfb4096aba66fed27554b", "submitted_at": "2026-01-28T23:38:25Z", "updated_at": "2026-02-02T23:11:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.21123v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.21123", "identifiers": {"arxiv": "2601.21123"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.21123"}], "name": "CUA-Skill: Develop Skills for Computer Using Agent", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "CUA-Skill encodes Windows application knowledge in parameterized skills and execution/composition graphs. Its agent retrieves and instantiates these skills and uses memory for failure recovery, with end-to-end evaluation on WindowsAgentArena.", "topics": ["tool-integration", "planning", "memory", "automation"], "updated_at": "2026-02-02", "url": "https://arxiv.org/abs/2601.21123"}
{"aliases": [], "authors": ["Xiang Zheng", "Yutao Wu", "Hanxun Huang", "Yige Li", "Xingjun Ma", "Bo Li", "Yu-Gang Jiang", "Cong Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7089ab1bfe278ba37ad21b11a0502d1a9c88111481e63d366b3ccd0073030d98", "submitted_at": "2026-01-29T03:53:25Z", "updated_at": "2026-06-26T10:38:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.21233v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.21233", "identifiers": {"arxiv": "2601.21233"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.21233"}], "name": "Just Ask: Curious Code Agents Reveal System Prompts in Frontier LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-26", "url": "https://arxiv.org/abs/2601.21233"}
{"aliases": [], "authors": ["Wei Minn", "Biniam Fisseha Demissie", "Yan Naing Tun", "Jiakun Liu", "Mariano Ceccato", "Lwin Khin Shar", "David Lo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac3f148047f7e63a33780fed3fd1ca461bead2128914c46f353efab4675df444", "submitted_at": "2026-01-29T04:21:11Z", "updated_at": "2026-01-29T04:21:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.21253v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.21253", "identifiers": {"arxiv": "2601.21253"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.21253"}], "name": "CovAgent: Overcoming the 30% Curse of Mobile Application Coverage with Agentic AI and Dynamic Instrumentation", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-01-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "CovAgent augments Android GUI fuzzers by analyzing decompiled app code and component transitions to identify unsatisfied activation conditions. A second agent generates instrumentation that enables otherwise difficult-to-reach activities, with coverage-based evaluation.", "topics": ["automation", "mobile-navigation", "evaluation"], "updated_at": "2026-01-29", "url": "https://arxiv.org/abs/2601.21253"}
{"aliases": [], "authors": ["Ziyu Lu", "Tengjin Weng", "Yiying Yang", "Yuhang Zhao", "Xinxin Huang", "Wenhao Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa36a9fbd0b189717b2a5311f06b6127d54f2eca09961cedd2a0b9d2a78b1b8b", "submitted_at": "2026-01-29T07:22:50Z", "updated_at": "2026-01-29T07:22:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.21352v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.21352", "identifiers": {"arxiv": "2601.21352"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.21352"}], "name": "BEAP-Agent: Backtrackable Execution and Adaptive Planning for GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "BEAP-Agent treats GUI execution as depth-first search with coordinated planning, execution and task-state tracking. It supports backtracking across multiple levels of earlier states and evaluates recovery on OSWorld.", "topics": ["planning", "state-tracking", "robustness", "long-horizon"], "updated_at": "2026-01-29", "url": "https://arxiv.org/abs/2601.21352"}
{"aliases": [], "authors": ["Xiang Li", "Ning Yan", "Masood Mortazavi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7bfb69c1569e80d8505e91f2083151076c032f0c27624341655c8aa2189cb411", "submitted_at": "2026-01-29T15:18:58Z", "updated_at": "2026-05-17T22:32:46Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.21841v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2601.21841", "identifiers": {"arxiv": "2601.21841"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.21841"}], "name": "Embodied Task Planning via Graph-Informed Action Generation with Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-17", "url": "https://arxiv.org/abs/2601.21841"}
{"aliases": [], "authors": ["Yao Zhang", "Shijie Tang", "Zeyu Li", "Zhen Han", "Volker Tresp"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68ae67856eb871d9a8f259dfb180bde02d29ee97fd7abf0e60719fc2c7abd7ab", "submitted_at": "2026-01-29T15:39:50Z", "updated_at": "2026-04-09T17:48:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.21872v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.21872", "identifiers": {"arxiv": "2601.21872"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.21872"}], "name": "WebArbiter: A Principle-Guided Reasoning Process Reward Model for Web Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebArbiter generates reasoned action-preference verdicts for web navigation, using reasoning distillation followed by reinforcement learning. It introduces WebPRMBench and evaluates both preference judgments and reward-guided trajectory search.", "topics": ["reward-models", "reinforcement-learning", "evaluation", "web-navigation"], "updated_at": "2026-04-09", "url": "https://arxiv.org/abs/2601.21872"}
{"aliases": [], "authors": ["Kuai Yu", "Naicheng Yu", "Han Wang", "Rui Yang", "Huan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1380697158d3a91334427cf067f3f91a751bd89d31103edecfa68171b28b8d66", "submitted_at": "2026-01-29T16:40:15Z", "updated_at": "2026-02-27T16:47:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.21961v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.21961", "identifiers": {"arxiv": "2601.21961"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.21961"}], "name": "How do Visual Attributes Influence Web Agents? A Comprehensive Evaluation of User Interface Design Factors", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-01-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces VAF, a controlled pipeline that varies webpage appearance while retaining item semantics. It measures both agent clicks and reasoning references to study which visual design factors change browsing decisions.", "topics": ["ui-design", "evaluation", "web-navigation"], "updated_at": "2026-02-27", "url": "https://arxiv.org/abs/2601.21961"}
{"aliases": [], "authors": ["Sami Abuzakuk", "Anne-Marie Kermarrec", "Rishi Sharma", "Rasmus Moorits Veski", "Martijn de Vos"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2fb6e25498e52e75e52c468cc499769cfdc17548584237c8411f9f90d950213d", "submitted_at": "2026-01-29T17:43:08Z", "updated_at": "2026-02-02T14:41:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.22037v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.22037", "identifiers": {"arxiv": "2601.22037"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.22037"}], "name": "Optimizing Agentic Workflows using Meta-tools", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-02", "url": "https://arxiv.org/abs/2601.22037"}
{"aliases": [], "authors": ["Hang Ding", "Peidong Liu", "Junqiao Wang", "Ziwei Ji", "Meng Cao", "Rongzhao Zhang", "Lynn Ai", "Eric Yang", "Tianyu Shi", "Lei Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aedb5ed90c91d72f62f5ab60be5374b6f431f887425bc760883ce3fb85a46d43", "submitted_at": "2026-01-29T18:59:07Z", "updated_at": "2026-04-17T18:27:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.22149v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.22149", "identifiers": {"arxiv": "2601.22149"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.22149"}], "name": "DynaWeb: Model-Based Reinforcement Learning of Web Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "DynaWeb learns a world model that predicts web-page representations after actions and uses it to generate policy-training rollouts. It interleaves these rollouts with expert trajectories and evaluates the trained web agents on WebArena and WebVoyager.", "topics": ["world-models", "reinforcement-learning", "data-generation", "web-navigation"], "updated_at": "2026-04-17", "url": "https://arxiv.org/abs/2601.22149"}
{"aliases": [], "authors": ["Kaixuan Fan", "Kaituo Feng", "Manyuan Zhang", "Tianshuo Peng", "Zhixun Li", "Yilei Jiang", "Shuang Chen", "Peng Pei", "Xunliang Cai", "Xiangyu Yue"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "460595bff5d0322a764b4aa4806f373722d03318659baaa8fa106d7baaa36a65", "submitted_at": "2026-01-29T18:59:52Z", "updated_at": "2026-04-28T11:53:16Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.22154v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.22154", "identifiers": {"arxiv": "2601.22154"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.22154"}], "name": "Exploring Reasoning Reward Model for Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-28", "url": "https://arxiv.org/abs/2601.22154"}
{"aliases": [], "authors": ["Jinyang Wu", "Changpeng Yang", "Yuhao Shen", "Fangzhi Xu", "Bolin Ni", "Chonghua Liao", "Yuchen Liu", "Hongzhen Wang", "Shuai Nie", "Shuai Zhang", "Haoran Luo", "Jiaming Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8105f329c0368b50d94091ee978791bf4c205ae124bb06cde504165b6b3066a", "submitted_at": "2026-01-30T03:02:18Z", "updated_at": "2026-01-30T03:02:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.22491v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2601.22491", "identifiers": {"arxiv": "2601.22491"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.22491"}], "name": "SSL: Sweet Spot Learning for Differentiated Guidance in Agentic Optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-30", "url": "https://arxiv.org/abs/2601.22491"}
{"aliases": [], "authors": ["Yuanjie Lyu", "Chengyu Wang", "Lei Shen", "Jun Huang", "Tong Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6de610858e1b5da453e0cd0ca261854a0f7d9f2dc720c36617bd01b47f859d0f", "submitted_at": "2026-01-30T03:43:42Z", "updated_at": "2026-03-12T08:14:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.22511v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.22511", "identifiers": {"arxiv": "2601.22511"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.22511"}], "name": "Mock Worlds, Real Skills: Building Small Agentic Language Models with Synthetic Tasks, Simulated Environments, and Rubric-Based Rewards", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-12", "url": "https://arxiv.org/abs/2601.22511"}
{"aliases": [], "authors": ["Hongze Mi", "Yibo Feng", "WenJie Lu", "Song Cao", "Jinyuan Li", "Yanming Li", "Xuelin Zhang", "Haotian Luo", "Songyang Peng", "He Cui", "Tengfei Tian", "Jun Fang", "Hua Chai", "Naiqiang Tan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0b3dbaacdf640a1314a40688133f7386c0eb8b901c43393ad21ec887814a606", "submitted_at": "2026-01-30T04:01:21Z", "updated_at": "2026-01-30T04:01:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.22528v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.22528", "identifiers": {"arxiv": "2601.22528"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.22528"}], "name": "Darwinian Memory: A Training-Free Self-Regulating Memory System for GUI Agent Evolution", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Darwinian Memory splits GUI trajectories into reusable units and updates the memory by retaining useful paths and pruning poor ones. It evaluates this training-free memory mechanism on long multi-application tasks.", "topics": ["memory", "training-free", "cross-app", "long-horizon"], "updated_at": "2026-01-30", "url": "https://arxiv.org/abs/2601.22528"}
{"aliases": [], "authors": ["Jiaxuan Gao", "Jiaao Chen", "Chuyi He", "Shusheng Xu", "Di Jin", "Yi Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14713aa02568608d82f7e2f33d9d550b069de2f002697b92d4ddede78cee57f6", "submitted_at": "2026-01-30T06:01:23Z", "updated_at": "2026-03-10T06:07:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.22607v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2601.22607", "identifiers": {"arxiv": "2601.22607"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2601.22607"}], "name": "From Self-Evolving Synthetic Data to Verifiable-Reward RL: Post-Training Multi-turn Interactive Tool-Using Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-10", "url": "https://arxiv.org/abs/2601.22607"}
{"aliases": [], "authors": ["Linjia Kang", "Zhimin Wang", "Yongkang Zhang", "Duo Wu", "Jinghe Wang", "Ming Ma", "Haopeng Yan", "Zhi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d66365682d0ea4cee744c6c59d3a585c9272a31da5f8ff7c27983d719f070a7", "submitted_at": "2026-01-30T10:03:20Z", "updated_at": "2026-01-30T10:03:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.22781v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2601.22781", "identifiers": {"arxiv": "2601.22781"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.22781"}], "name": "Learning with Challenges: Adaptive Difficulty-Aware Data Generation for Mobile GUI Agent Training", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileGen estimates a mobile agent’s strengths across structural and semantic task difficulty, then samples new training targets around that profile. A controllable multi-agent generator creates the corresponding instructions and GUI trajectories.", "topics": ["data-generation", "mobile-navigation", "evaluation"], "updated_at": "2026-01-30", "url": "https://arxiv.org/abs/2601.22781"}
{"aliases": [], "authors": ["Chaoqun Cui", "Jing Huang", "Shijing Wang", "Liming Zheng", "Qingchao Kong", "Zhixiong Zeng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86225f2e52e1394e2a0a56b1e82327d756bff95739e01ce7cc5898af49944aba", "submitted_at": "2026-01-31T07:36:54Z", "updated_at": "2026-07-25T06:54:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.00575v2"}}, "environments": ["desktop", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.00575", "identifiers": {"arxiv": "2602.00575"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.00575"}], "name": "Agentic Reward Modeling: Verifying GUI Agent via Progressive Trajectory-Grounded Interaction", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "VAGEN verifies GUI trajectories by first extracting readily available evidence and escalating to tool-based environment probes when needed. Evaluation on OSWorld-Verified and AndroidWorld studies verification accuracy and its cost.", "topics": ["reward-models", "tool-integration", "evaluation", "efficiency"], "updated_at": "2026-07-25", "url": "https://arxiv.org/abs/2602.00575"}
{"aliases": [], "authors": ["Alberto Castelo", "Zahra Zanjani Foumani", "Ailin Fan", "Keat Yang Koay", "Vibhor Malik", "Yuanzheng Zhu", "Han Li", "Meysam Feghhi", "Ronie Uliana", "Shuang Xie", "Zhaoyu Zhang", "Angelo Ocana Martins", "Mingyu Zhao", "Francis Pelland", "Jonathan Faerman", "Nikolas LeBlanc", "Aaron Glazer", "Andrew McNamara", "Lingyun Wang", "Zhong Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6276599a10318323b7845753a0e0a3eecf518fbec32e3355d7ace6bff796660b", "submitted_at": "2026-02-01T21:23:04Z", "updated_at": "2026-02-01T21:23:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.01443v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.01443", "identifiers": {"arxiv": "2602.01443"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.01443"}], "name": "SimGym: Traffic-Grounded Browser Agents for Offline A/B Testing in E-Commerce", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-01", "url": "https://arxiv.org/abs/2602.01443"}
{"aliases": [], "authors": ["S1-NexusAgent Team"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "867a6bf56636fde5d034cd6b4dbbc54668d309ff4fb6981466bc9854defd37e6", "submitted_at": "2026-02-02T02:33:25Z", "updated_at": "2026-02-10T10:14:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.01550v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.01550", "identifiers": {"arxiv": "2602.01550"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.01550"}], "name": "S1-NexusAgent: a Self-Evolving Agent Framework for Multidisciplinary Scientific Research", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-10", "url": "https://arxiv.org/abs/2602.01550"}
{"aliases": [], "authors": ["Woosung Koh", "Sungjun Han", "Segyu Lee", "Se-Young Yun", "Jamin Shin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c29d1404b02c4f5ef2cdaf78bdb0f9062b1b85ec1a154408bc8f49446c28d08", "submitted_at": "2026-02-02T03:12:16Z", "updated_at": "2026-05-23T02:41:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.01576v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2602.01576", "identifiers": {"arxiv": "2602.01576"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.01576"}], "name": "Generative Visual Code Mobile World Models", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "gWorld predicts future mobile GUI states as renderable web code and supplies an automated pipeline for code-based training data. The study evaluates visual state prediction and how stronger world models affect downstream mobile policies.", "topics": ["world-models", "data-generation", "screen-understanding", "mobile-navigation"], "updated_at": "2026-05-23", "url": "https://arxiv.org/abs/2602.01576"}
{"aliases": [], "authors": ["Yurun Chen", "Zeyi Liao", "Ping Yin", "Taotao Xie", "Keting Yin", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d1a6f98c0ab0a221c169796d9680e4b2a7bf7a0775c0493254f47d5e0a9b3c4e", "submitted_at": "2026-02-02T07:04:06Z", "updated_at": "2026-02-02T07:04:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.01725v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:38+00:00", "role": "primary", "sha256": "6e1266c5a47cd667d681d3143a0331b74b6f64416b7e01f44a93a001c38ef74b", "supports": ["summary", "environments", "scope"], "url": "https://arxiv.org/html/2602.01725v1"}], "id": "arxiv:2602.01725", "identifiers": {"arxiv": "2602.01725"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.01725"}], "name": "SafePred: A Predictive Guardrail for Computer-Using Agents via World Models", "notes": "The annotation was reviewed against the complete retained official abstract and the specified sections of the same recorded paper version. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "SafePred uses world-model risk forecasts to guide action filtering and task replanning. Its desktop and web safety evaluations distinguish policy compliance, task success and success under the specified policy.", "topics": ["safety-security", "world-models", "planning", "evaluation"], "updated_at": "2026-02-02", "url": "https://arxiv.org/abs/2602.01725"}
{"aliases": [], "authors": ["Qingni Wang", "Yue Fan", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c9557fc608a7fd7667746e9b3e06bc699b1796cdd77fe5e975ba1bd5a64690c1", "submitted_at": "2026-02-02T18:22:45Z", "updated_at": "2026-02-03T08:31:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.02419v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.02419", "identifiers": {"arxiv": "2602.02419"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02419"}], "name": "SafeGround: Know When to Trust GUI Grounding Models via Uncertainty Calibration", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SafeGround, which estimates GUI-grounding uncertainty from the spatial dispersion of sampled predictions and calibrates a decision threshold before testing. It studies risk-aware selection on ScreenSpot-Pro, with statistical control scoped to the threshold’s false-discovery rate.", "topics": ["grounding", "robustness", "evaluation"], "updated_at": "2026-02-03", "url": "https://arxiv.org/abs/2602.02419"}
{"aliases": [], "authors": ["Aiden Yiliu Li", "Xinyue Hao", "Shilong Liu", "Mengdi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5b730fa81936d6a1b2b0a8b74a522a22de39b977929165c44a5050e490c301f", "submitted_at": "2026-02-02T18:50:07Z", "updated_at": "2026-02-02T18:50:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02468v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.02468", "identifiers": {"arxiv": "2602.02468"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02468"}], "name": "Avenir-Web: Human-Experience-Imitating Multimodal Web Agents with Mixture of Grounding Experts", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Avenir-Web combines multiple grounding specialists with procedural experience, a task checklist and adaptive memory. It evaluates this planning and execution framework on live Online-Mind2Web tasks.", "topics": ["grounding", "planning", "memory", "web-navigation"], "updated_at": "2026-02-02", "url": "https://arxiv.org/abs/2602.02468"}
{"aliases": [], "authors": ["Xiaoce Wang", "Guibin Zhang", "Junzhe Li", "Jinzhe Tu", "Chun Li", "Ming Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49b0dbff31dcf8c966bebd367ae05984e46a7fd13b2cfb1ad5ef061332b43f00", "submitted_at": "2026-01-30T08:38:05Z", "updated_at": "2026-01-30T08:38:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02548v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.02548", "identifiers": {"arxiv": "2602.02548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02548"}], "name": "ToolTok: Tool Tokenization for Efficient and Generalizable GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-01-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "ToolTok represents GUI operations with learned tool tokens and frames interaction as progressive tool selection. Semantic anchoring and an easy-to-hard curriculum teach tool meanings before visual pathfinding, with evaluation across grounding scenarios.", "topics": ["tool-integration", "grounding", "data-generation"], "updated_at": "2026-01-30", "url": "https://arxiv.org/abs/2602.02548"}
{"aliases": [], "authors": ["Sizhe Tang", "Rongqian Chen", "Tian Lan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a01499ce704bb7916a3157ff2d3fc755a1115d9b27c7e529c153bce360a5404", "submitted_at": "2026-02-03T02:07:12Z", "updated_at": "2026-02-03T02:07:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02995v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.02995", "identifiers": {"arxiv": "2602.02995"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02995"}], "name": "Agent Alpha: Tree Search Unifying Generation, Exploration and Evaluation for Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Agent Alpha uses step-level Monte Carlo tree search to reuse successful prefixes and recover from early GUI mistakes. Comparison-based evaluation and diversity-constrained expansion guide the search, with experiments on OSWorld.", "topics": ["planning", "efficiency", "evaluation"], "updated_at": "2026-02-03", "url": "https://arxiv.org/abs/2602.02995"}
{"aliases": [], "authors": ["Hao Li", "Ruoyao Wen", "Shanghao Shi", "Ning Zhang", "Yevgeniy Vorobeychik", "Chaowei Xiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7db3fca35b801ffab4d1ab4dfc84122516270894766f2cf1f641d2bdd3ce9216", "submitted_at": "2026-02-03T05:20:42Z", "updated_at": "2026-05-07T07:42:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.03117v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.03117", "identifiers": {"arxiv": "2602.03117"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.03117"}], "name": "AgentDyn: Are Your Agent Security Defenses Deployable in Real-World Dynamic Environments?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-07", "url": "https://arxiv.org/abs/2602.03117"}
{"aliases": [], "authors": ["Tianyu Chen", "Chujia Hu", "Dongrui Liu", "Xia Hu", "Wenjie Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e35743f909d7649d08d8b154486a49c0393149ba0e24a2c4bdea519079da6359", "submitted_at": "2026-02-03T08:40:24Z", "updated_at": "2026-10-02T00:39:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.03255v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.03255", "identifiers": {"arxiv": "2602.03255"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.03255"}], "name": "LPS-Bench: Benchmarking Safety Awareness of Computer-Use Agents in Long-Horizon Planning under Benign and Adversarial Scenarios", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-02", "url": "https://arxiv.org/abs/2602.03255"}
{"aliases": [], "authors": ["Xilong Wang", "Yinuo Liu", "Zhun Wang", "Dawn Song", "Neil Gong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86702018906e64294d7b8daea618c5f152e8b94b604d27d6b164e82aff462564", "submitted_at": "2026-02-03T17:55:04Z", "updated_at": "2026-02-03T17:55:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.03792v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.03792", "identifiers": {"arxiv": "2602.03792"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.03792"}], "name": "WebSentinel: Detecting and Localizing Prompt Injection Attacks for Web Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebSentinel identifies suspicious webpage segments and checks each against its surrounding page context to detect and locate prompt injections. Its evaluation uses both contaminated and clean webpage datasets.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2026-02-03", "url": "https://arxiv.org/abs/2602.03792"}
{"aliases": [], "authors": ["Syed Nazmus Sakib", "Nafiul Haque", "Tapodhir Karmakar Taton", "Shahrear Bin Amin", "Shifat E. Arman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76fbeae086fb9ef21f820b02777a2ceb66ee37f8f1c978c7426589ea4e888e49", "submitted_at": "2026-02-05T06:24:23Z", "updated_at": "2026-09-08T14:40:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.05354v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.05354", "identifiers": {"arxiv": "2602.05354"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.05354"}], "name": "Do Web Agents Investigate Before They Decide?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MIRAGE, a web decision benchmark where agents must actively uncover evidence that can contradict the visible context. Separate investigation, reasoning, decision, and hallucination measures diagnose failures beyond reaching the right page.", "topics": ["evaluation", "web-navigation", "planning"], "updated_at": "2026-09-08", "url": "https://arxiv.org/abs/2602.05354"}
{"aliases": [], "authors": ["Rui Lv", "Juncheng Mo", "Tianyi Chu", "Chen Rao", "Hongyi Jing", "Jiajie Teng", "Jiafu Chen", "Shiqi Zhang", "Liangzi Ding", "Shuo Fang", "Huaizhong Lin", "Ziqiang Dang", "Chenguang Ma", "Lei Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "84d822655c810958cff6035584542d7e151d765734ec7030f841c15561ee5677", "submitted_at": "2026-02-05T08:19:39Z", "updated_at": "2026-02-05T08:19:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.05429v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.05429", "identifiers": {"arxiv": "2602.05429"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.05429"}], "name": "M$^2$-Miner: Multi-Agent Enhanced MCTS for Mobile GUI Agent Data Mining", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "M2-Miner uses multi-agent-guided Monte Carlo tree search to collect mobile GUI intent–trajectory pairs. Intent recycling and progressive collector training aim to increase the diversity and usefulness of the mined interaction data.", "topics": ["data-generation", "mobile-navigation", "planning"], "updated_at": "2026-02-05", "url": "https://arxiv.org/abs/2602.05429"}
{"aliases": [], "authors": ["Chang Yang", "Chuang Zhou", "Yilin Xiao", "Su Dong", "Luyao Zhuang", "Yujing Zhang", "Zhu Wang", "Zijin Hong", "Zheng Yuan", "Zhishang Xiang", "Shengyuan Chen", "Huachi Zhou", "Qinggang Zhang", "Ninghao Liu", "Jinsong Su", "Xinrun Wang", "Yi Chang", "Xiao Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77ba8dae47d8164b5ad86c6da98f2678a035c05ef91b68fe40a8d7f49002c950", "submitted_at": "2026-02-05T13:49:05Z", "updated_at": "2026-02-05T13:49:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.05665v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.05665", "identifiers": {"arxiv": "2602.05665"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.05665"}], "name": "Graph-based Agent Memory: Taxonomy, Techniques, and Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-05", "url": "https://arxiv.org/abs/2602.05665"}
{"aliases": [], "authors": ["Xin Chen", "Jie Zhang", "Florian Tramèr"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "19dde6dc1fe2c61cc10393aaef02a226b49f12015275db6f276c9b6ea33746fa", "submitted_at": "2026-02-05T15:14:46Z", "updated_at": "2026-06-10T08:55:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.05746v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.05746", "identifiers": {"arxiv": "2602.05746"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.05746"}], "name": "Learning to Inject: Automated Prompt Injection via Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-10", "url": "https://arxiv.org/abs/2602.05746"}
{"aliases": [], "authors": ["Han Xiao", "Guozhi Wang", "Hao Wang", "Shilong Liu", "Yuxiang Chai", "Yue Pan", "Yufeng Zhou", "Xiaoxin Chen", "Yafei Wen", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f51eb0b7102a37fcc705c53640c5a1d30fcc115cedcdca4a00493ed45b9d5507", "submitted_at": "2026-02-05T16:21:43Z", "updated_at": "2026-02-05T16:21:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.05832v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.05832", "identifiers": {"arxiv": "2602.05832"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.05832"}], "name": "UI-Mem: Self-Evolving Experience Memory for Online Reinforcement Learning in Mobile GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-Mem stores workflows, skills and failure patterns as reusable templates to guide online GUI reinforcement learning. It varies guidance within rollout groups and continually updates the memory so an unguided policy can learn from guided experiences.", "topics": ["memory", "reinforcement-learning", "cross-app"], "updated_at": "2026-02-05", "url": "https://arxiv.org/abs/2602.05832"}
{"aliases": [], "authors": ["Lukas Stappen", "Ahmet Erkan Turan", "Johann Hagerer", "Georg Groh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b14fa6beb8450f133f64130158cf71a8e286ba5522af8abe4e02fc27c4f1d0e", "submitted_at": "2026-02-05T16:53:41Z", "updated_at": "2026-02-05T16:53:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.05877v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.05877", "identifiers": {"arxiv": "2602.05877"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.05877"}], "name": "Agent2Agent Threats in Safety-Critical LLM Assistants: A Human-Centric Taxonomy", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-05", "url": "https://arxiv.org/abs/2602.05877"}
{"aliases": [], "authors": ["Wei-Chieh Huang", "Weizhi Zhang", "Yueqing Liang", "Yuanchen Bei", "Yankai Chen", "Tao Feng", "Xinyu Pan", "Zhen Tan", "Yu Wang", "Tianxin Wei", "Shanglin Wu", "Ruiyao Xu", "Liangwei Yang", "Rui Yang", "Wooseong Yang", "Chin-Yuan Yeh", "Hanrong Zhang", "Haozhen Zhang", "Siqi Zhu", "Henry Peng Zou", "Wanjia Zhao", "Song Wang", "Wujiang Xu", "Zixuan Ke", "Zheng Hui", "Dawei Li", "Yaozu Wu", "Langzhou He", "Chen Wang", "Xiongxiao Xu", "Baixiang Huang", "Juntao Tan", "Shelby Heinecke", "Huan Wang", "Caiming Xiong", "Ahmed A. Metwally", "Jun Yan", "Chen-Yu Lee", "Hanqing Zeng", "Yinglong Xia", "Xiaokai Wei", "Ali Payani", "Yu Wang", "Haitong Ma", "Wenya Wang", "Chenguang Wang", "Yu Zhang", "Xin Eric Wang", "Yongfeng Zhang", "Jiaxuan You", "Hanghang Tong", "Xiao Luo", "Xue Liu", "Yizhou Sun", "Wei Wang", "Julian McAuley", "James Zou", "Jiawei Han", "Philip S. Yu", "Kai Shu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0be0b413b7774a44e08c92ae75b3a017976d693048a095137950f32164a80df8", "submitted_at": "2026-01-14T07:38:38Z", "updated_at": "2026-08-04T20:54:45Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2602.06052v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.06052", "identifiers": {"arxiv": "2602.06052"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.06052"}], "name": "A Survey of Agent Memory in the Second Half: Towards Self-Evolving and Long-Horizon Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2602.06052"}
{"aliases": [], "authors": ["Guangyi Liu", "Pengxiang Zhao", "Yaozhen Liang", "Qinyi Luo", "Shunye Tang", "Yuxiang Chai", "Weifeng Lin", "Han Xiao", "WenHao Wang", "Siheng Chen", "Zhengxi Lu", "Gao Wu", "Hao Wang", "Liang Liu", "Yong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0692d6936a3e3c20a7ad61c69ae770f5669c3783ff837092254f8b3fb170cd0a", "submitted_at": "2026-02-03T17:01:59Z", "updated_at": "2026-08-27T06:48:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.06075v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.06075", "identifiers": {"arxiv": "2602.06075"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.06075"}], "name": "MemGUI-Bench: Benchmarking Memory of Mobile GUI Agents in Dynamic Environments", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "MemGUI-Bench evaluates short-term retention and cross-session experience reuse in mobile GUI agents using paired tasks and repeated-attempt protocols. Its staged evaluator measures memory fidelity, learning and execution efficiency.", "topics": ["memory", "evaluation", "mobile-navigation", "cross-app"], "updated_at": "2026-08-27", "url": "https://arxiv.org/abs/2602.06075"}
{"aliases": [], "authors": ["Jincheng Wang", "Liwei Luo", "Xiaoxuan Tang", "Jingxuan Xu", "Sheng Zhou", "Dajun Chen", "Wei Jiang", "Yong Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccb28161c4665de8ff69d7142c91d5531b77350a04df5692dfc9fcfc8744814b", "submitted_at": "2026-02-05T06:26:46Z", "updated_at": "2026-08-03T12:05:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.06090v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "id": "arxiv:2602.06090", "identifiers": {"arxiv": "2602.06090"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.06090"}], "name": "SVRepair: Structured Visual Reasoning for Automated Program Repair", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-03", "url": "https://arxiv.org/abs/2602.06090"}
{"aliases": [], "authors": ["Longhui Ma", "Di Zhao", "Siwei Wang", "Zhao Lv", "Miao Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2591e0dab6a84b5db2631ddc94a2c0b65b13a97d06d411ab6ad4e243fe3d3e1e", "submitted_at": "2026-02-06T03:27:55Z", "updated_at": "2026-02-06T03:27:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.06351v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.06351", "identifiers": {"arxiv": "2602.06351"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.06351"}], "name": "Trifuse: Enhancing Attention-Based GUI Grounding via Multimodal Fusion", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trifuse combines attention-based localization with OCR text cues and icon captions to ground GUI instructions. Its fusion procedure favors cross-modal agreement and sharp location peaks, and is evaluated without task-specific fine-tuning.", "topics": ["grounding", "screen-understanding", "training-free"], "updated_at": "2026-02-06", "url": "https://arxiv.org/abs/2602.06351"}
{"aliases": [], "authors": ["Zhongyin Zhao", "Yuan Liu", "Yikun Liu", "Haicheng Wang", "Le Tian", "Xiao Zhou", "Yangxiu You", "Zilin Yu", "Yang Yu", "Jie Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "852813305299178baeda3540cbad539dc30b840681e7bbdaed34dac1341dc0de", "submitted_at": "2026-02-06T05:14:11Z", "updated_at": "2026-02-06T05:14:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.06391v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.06391", "identifiers": {"arxiv": "2602.06391"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.06391"}], "name": "POINTS-GUI-G: GUI-Grounding Journey", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies GUI-grounding development from a weakly grounded base model through data preparation, visual training, and verifiable rewards.", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2026-02-06", "url": "https://arxiv.org/abs/2602.06391"}
{"aliases": [], "authors": ["Junxian Li", "Kai Liu", "Leyang Chen", "Weida Wang", "Zhixin Wang", "Jiaqi Xu", "Fan Li", "Renjing Pei", "Linghe Kong", "Yulun Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e5678170e8db3f727f0fbae211f6cc9c2bdc6d748331ee964e2f12a34730d1aa", "submitted_at": "2026-02-06T12:47:16Z", "updated_at": "2026-04-20T06:42:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.06663v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.06663", "identifiers": {"arxiv": "2602.06663"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.06663"}], "name": "PlanViz: Evaluating Planning-Oriented Image Generation and Editing for Computer-Use Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-20", "url": "https://arxiv.org/abs/2602.06663"}
{"aliases": [], "authors": ["Ruoyu Chen", "Shangquan Sun", "Xiaoqing Guo", "Kangwei Liu", "Sanyi Zhang", "Zhangcheng Wang", "Shiming Liu", "Qunli Zhang", "Hua Zhang", "Wei Wang", "Xiaochun Cao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d517d1cbd3bb6c7e835d777d6ed5211cf9f0eca7596a7acf8fbdff1aaae14ef4", "submitted_at": "2026-01-30T10:29:27Z", "updated_at": "2026-09-26T09:11:54Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2602.07008v6"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2602.07008", "identifiers": {"arxiv": "2602.07008"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.07008"}], "name": "Where Not to Learn: Prior-Aligned Training with Subset-based Attribution Constraints", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2602.07008"}
{"aliases": [], "authors": ["Jinbiao Wei", "Yilun Zhao", "Kangqi Ni", "Arman Cohan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce896d4d846bf07fedc68b7f952532c4fefe24dcb49b51556e0d6c2bc648bb91", "submitted_at": "2026-02-06T19:55:26Z", "updated_at": "2026-04-12T06:09:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.07153v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.07153", "identifiers": {"arxiv": "2602.07153"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.07153"}], "name": "ANCHOR: Branch-Point Data Generation for GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "ANCHOR expands verified desktop demonstrations by generating new tasks at meaningful branch points and executing their instructions. State-aware verification and action filtering refine the resulting trajectories, which are evaluated through fine-tuned agents on OSWorld and WindowsAgentArena.", "topics": ["data-generation", "planning", "evaluation"], "updated_at": "2026-04-12", "url": "https://arxiv.org/abs/2602.07153"}
{"aliases": [], "authors": ["Ruijia Cheng", "Jenny T. Liang", "Eldon Schoop", "Jeffrey Nichols"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d556ebbfbf4fc5875395b449a92b8ff27a7ed59d41bc062ac60fcf175f9064f8", "submitted_at": "2026-02-07T00:19:08Z", "updated_at": "2026-02-07T00:19:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.07283v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.07283", "identifiers": {"arxiv": "2602.07283"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.07283"}], "name": "Mapping the Design Space of User Experience for Computer Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Develops a taxonomy of computer-use-agent interaction design through system review and practitioner interviews. A Wizard-of-Oz study then examines how users respond to normal, erroneous, and risky web-based execution, focusing on control, explanation, prompts, and mental models.", "topics": ["human-in-the-loop", "ui-design", "evaluation"], "updated_at": "2026-02-07", "url": "https://arxiv.org/abs/2602.07283"}
{"aliases": [], "authors": ["Ruoyao Wen", "Hao Li", "Chaowei Xiao", "Ning Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ab315309ef09e888cf18db7327f42a2f32171adad70bc835f8d00456d57518c9", "submitted_at": "2026-02-07T06:28:51Z", "updated_at": "2026-02-07T06:28:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.07398v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.07398", "identifiers": {"arxiv": "2602.07398"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.07398"}], "name": "AgentSys: Secure and Dynamic LLM Agents Through Explicit Hierarchical Memory Management", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-07", "url": "https://arxiv.org/abs/2602.07398"}
{"aliases": [], "authors": ["Yancheng Long", "Yankai Yang", "Hongyang Wei", "Wei Chen", "Tianke Zhang", "Haonan fan", "Changyi Liu", "Kaiyu Jiang", "Jiankang Chen", "Kaiyu Tang", "Bin Wen", "Fan Yang", "Tingting Gao", "Han Li", "Shuo Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff3289631105e3b73ed51024a967a1665578b9c06d14ce93bc79f3ed9d1f5646", "submitted_at": "2026-02-07T09:23:34Z", "updated_at": "2026-05-13T14:58:33Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2602.07458v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.07458", "identifiers": {"arxiv": "2602.07458"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.07458"}], "name": "SpatialReward: Bridging the Perception Gap in Online RL for Image Editing via Explicit Spatial Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-13", "url": "https://arxiv.org/abs/2602.07458"}
{"aliases": [], "authors": ["Junyu Feng", "Binxiao Xu", "Jiayi Chen", "Mengyu Dai", "Cenyang Wu", "Haodong Li", "Bohan Zeng", "Yunliu Xie", "Hao Liang", "Ming Lu", "Wentao Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f39306d72f1b0b04cf4c43e068009fae45e92f34309ae88dc2a5b9a77b37fddd", "submitted_at": "2026-02-07T17:13:56Z", "updated_at": "2026-02-07T17:13:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.07624v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2602.07624", "identifiers": {"arxiv": "2602.07624"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.07624"}], "name": "M2A: Multimodal Memory Agent with Dual-Layer Hybrid Memory for Long-Term Personalized Interactions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-07", "url": "https://arxiv.org/abs/2602.07624"}
{"aliases": [], "authors": ["Pierre-Louis Favreau", "Jean-Pierre Lo", "Clement Guiguet", "Charles Simon-Meunier", "Nicolas Dehandschoewercker", "Allen G. Roush", "Judah Goldfeder", "Ravid Shwartz-Ziv"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ca42c9befea71d930c4b4f38a5eb4f6fb812a4966623556a56d417e3dec229af", "submitted_at": "2026-02-08T03:02:13Z", "updated_at": "2026-02-08T03:02:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.07787v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2602.07787", "identifiers": {"arxiv": "2602.07787"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.07787"}], "name": "Do Multi-Agents Dream of Electric Screens? Achieving Perfect Accuracy on AndroidWorld Through Task Decomposition", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-08", "url": "https://arxiv.org/abs/2602.07787"}
{"aliases": [], "authors": ["Minbeom Kim", "Mihir Parmar", "Phillip Wallis", "Lesly Miculicich", "Kyomin Jung", "Krishnamurthy Dj Dvijotham", "Long T. Le", "Tomas Pfister"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "864ede00f012511dcec356bc6ab040f28b831099a5cf92fb31c51c0ec1008d26", "submitted_at": "2026-02-08T11:34:08Z", "updated_at": "2026-02-08T11:34:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.07918v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.07918", "identifiers": {"arxiv": "2602.07918"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.07918"}], "name": "CausalArmor: Efficient Indirect Prompt Injection Guardrails via Causal Attribution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-08", "url": "https://arxiv.org/abs/2602.07918"}
{"aliases": [], "authors": ["Jaylen Jones", "Zhehao Zhang", "Yuting Ning", "Eric Fosler-Lussier", "Pierre-Luc St-Charles", "Yoshua Bengio", "Dawn Song", "Yu Su", "Huan Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d6e474c42c3e43b36530810134c68cf13491ceab43de3e2d5fcc0bc65587163", "submitted_at": "2026-02-09T03:20:11Z", "updated_at": "2026-06-06T21:46:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.08235v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.08235", "identifiers": {"arxiv": "2602.08235"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.08235"}], "name": "When Benign Inputs Lead to Severe Harms: Eliciting Unsafe Unintended Behaviors of Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-06", "url": "https://arxiv.org/abs/2602.08235"}
{"aliases": [], "authors": ["Yuting Ning", "Jaylen Jones", "Zhehao Zhang", "Chentao Ye", "Weitong Ruan", "Junyi Li", "Rahul Gupta", "Huan Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "853f64c020f47129be01778d5212ade8ef78251959ad634ee04ad0fb9fe2989c", "submitted_at": "2026-02-09T18:41:15Z", "updated_at": "2026-06-25T04:36:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.08995v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:26:41+00:00", "role": "primary", "sha256": "487a5d21d2385320315eb5a4a2c76e1017973d6027e219be6d31d9328fd5ff9a", "supports": ["summary", "environments", "scope"], "url": "https://arxiv.org/html/2602.08995v2"}], "id": "arxiv:2602.08995", "identifiers": {"arxiv": "2602.08995"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.08995"}], "name": "When Actions Go Off-Task: Detecting and Correcting Misaligned Actions in Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract and the specified sections of the same recorded paper version. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "MisActBench provides action-level labels for task alignment, while DeAction checks proposed computer-use actions and requests revisions through structured feedback. The paper separately evaluates offline detection and interactive operation in OSWorld and RedTeamCUA.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-06-25", "url": "https://arxiv.org/abs/2602.08995"}
{"aliases": [], "authors": ["Haodong Li", "Jingwei Wu", "Quan Sun", "Guopeng Li", "Juanxi Tian", "Ziyu Guo", "Donghao Zhou", "Tianfei Ren", "Huanyu Zhang", "Yanlin Lai", "Ruichuan An", "Hongbo Peng", "Yuhong Dai", "Jia Wang", "Ziyang Meng", "Zheng Ge", "Xiangyu Zhang", "Daxin Jiang", "Pheng-Ann Heng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43a01e0eb848a57a84767836d2c3c982bbf6094d22a3e70652dd42af52760bc5", "submitted_at": "2026-02-09T18:52:02Z", "updated_at": "2026-09-26T18:28:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.09007v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.09007", "identifiers": {"arxiv": "2602.09007"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09007"}], "name": "GUI-GenBench: Evaluating Image Generation Models as Interactive GUI Environments", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-02-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-GenBench tests image-generation models as action-conditioned GUI environments across five suites, including transitions, trajectories and grounding. Its VLM-based GUI-Score separates goal completion, interaction logic, consistency and visual properties.", "topics": ["evaluation", "world-models", "grounding", "screen-understanding"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2602.09007"}
{"aliases": [], "authors": ["Jiacheng Liu", "Yaxin Luo", "Jiacheng Cui", "Xinyi Shang", "Xiaohan Zhao", "Zhiqiang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3178684a322fcf07f847473139efdddf0b02213af468f1afbd285549d15bde56", "submitted_at": "2026-02-09T18:55:33Z", "updated_at": "2026-06-21T08:25:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.09012v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.09012", "identifiers": {"arxiv": "2602.09012"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09012"}], "name": "Next-Gen CAPTCHAs: Leveraging the Cognitive Gap for Scalable and Diverse GUI-Agent Defense", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Next-Gen CAPTCHAs introduces generated interactive challenges intended to distinguish human users from GUI agents. The benchmark targets perception, memory, decisions and actions, with backend-supported challenge types designed for scalable instance generation.", "topics": ["evaluation", "safety-security", "data-generation"], "updated_at": "2026-06-21", "url": "https://arxiv.org/abs/2602.09012"}
{"aliases": [], "authors": ["Venus Team", "Changlong Gao", "Zhangxuan Gu", "Yulin Liu", "Xinyu Qiu", "Shuheng Shen", "Yue Wen", "Tianyu Xia", "Zhenyu Xu", "Zhengwen Zeng", "Beitong Zhou", "Xingran Zhou", "Weizhi Chen", "Sunhao Dai", "Jingya Dou", "Yichen Gong", "Yuan Guo", "Zhenlin Guo", "Feng Li", "Qian Li", "Jinzhen Lin", "Yuqi Zhou", "Linchao Zhu", "Liang Chen", "Zhenyu Guo", "Changhua Meng", "Weiqiang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "efc4ad8f13d55939abcc783bdd652f276c3ac0164ea35232a3fb9366642d4558", "submitted_at": "2026-02-09T18:43:40Z", "updated_at": "2026-02-24T15:17:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.09082v2"}}, "environments": ["cross-platform", "web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2602.09082", "identifiers": {"arxiv": "2602.09082"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09082"}], "name": "UI-Venus-1.5 Technical Report", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a GUI-agent family combining broad intermediate training, full-trajectory online reinforcement learning, and merging of specialized models.", "topics": ["reinforcement-learning", "grounding", "mobile-navigation"], "updated_at": "2026-02-24", "url": "https://arxiv.org/abs/2602.09082"}
{"aliases": [], "authors": ["Georgios Syros", "Evan Rose", "Brian Grinstead", "Christoph Kerschbaumer", "William Robertson", "Cristina Nita-Rotaru", "Alina Oprea"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77fa6f48314afe90107775dac6c54b56570a7284f2b542e3048a590d87c8c118", "submitted_at": "2026-02-09T21:46:18Z", "updated_at": "2026-06-14T21:10:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.09222v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.09222", "identifiers": {"arxiv": "2602.09222"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.09222"}], "name": "MUZZLE: Adaptive Agentic Red-Teaming of Web Agents Against Indirect Prompt Injection Attacks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-14", "url": "https://arxiv.org/abs/2602.09222"}
{"aliases": [], "authors": ["Ananya Gubbi Mohanbabu", "Rosiana Natalie", "Brandon Kim", "Anhong Guo", "Amy Pavel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c3606d090749737d4223a9f826400c2b2aebde245c623391f2a8c6702b2468f8", "submitted_at": "2026-02-10T01:00:55Z", "updated_at": "2026-02-10T01:00:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.09310v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.09310", "identifiers": {"arxiv": "2602.09310"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09310"}], "name": "A11y-CUA Dataset: Characterizing the Accessibility Gap in Computer Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces A11y-CUA, a dataset comparing how blind or low-vision users and sighted users perform everyday computer tasks. It contrasts those traces with agent behavior under default, keyboard-only, and magnifier conditions to identify barriers to shared control.", "topics": ["human-in-the-loop", "evaluation", "data-generation"], "updated_at": "2026-02-10", "url": "https://arxiv.org/abs/2602.09310"}
{"aliases": [], "authors": ["Furong Jia", "Ling Dai", "Wenjin Deng", "Fan Zhang", "Chen Hu", "Daxin Jiang", "Yu Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03329722ce130df42c75d1322a30c0d3dfdababadeb9d7d9f7cc342e5bbef41a", "submitted_at": "2026-02-10T06:57:12Z", "updated_at": "2026-03-02T06:18:27Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.09463v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.09463", "identifiers": {"arxiv": "2602.09463"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.09463"}], "name": "SpotAgent: Grounding Visual Geo-localization in Large Vision-Language Models through Agentic Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-02", "url": "https://arxiv.org/abs/2602.09463"}
{"aliases": [], "authors": ["Xavier Hu", "Jinxiang Xia", "Shengze Xu", "Kangqi Song", "Yishuo Yuan", "Guibin Zhang", "JinCheng Ren", "Boyu Feng", "Li Lu", "Tieyong Zeng", "Jiaheng Liu", "Minghao Liu", "He Zhu", "Yuchen Eleanor Jiang", "Wei Wang", "Wangchunshu Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "724b7e95f1ab6ee2e9c5e8bb91b17cf4ec977df7f23f853d856d96312b2293b6", "submitted_at": "2026-02-10T08:12:23Z", "updated_at": "2026-05-09T05:05:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.09514v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.09514", "identifiers": {"arxiv": "2602.09514"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.09514"}], "name": "EcoGym: Evaluating LLMs for Long-Horizon Plan-and-Execute in Interactive Economies", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-09", "url": "https://arxiv.org/abs/2602.09514"}
{"aliases": [], "authors": ["Deyang Jiang", "Jing Huang", "Xuanle Zhao", "Lei Chen", "Liming Zheng", "Fanfan Liu", "Haibo Qiu", "Peng Shi", "Zhixiong Zeng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b5080050b85e38c8c03078c96c206e92d4ed63d106d55983563d51fcbc9f66c5", "submitted_at": "2026-02-10T11:16:57Z", "updated_at": "2026-02-10T11:16:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.09662v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2602.09662", "identifiers": {"arxiv": "2602.09662"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09662"}], "name": "TreeCUA: Efficiently Scaling GUI Automation with Tree-Structured Verifiable Evolution", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "TreeCUA organizes GUI exploration trajectories as reusable trees and balances deeper tasks against broader coverage. Its data pipeline combines action verification, memory-backed recovery and a DPO variant that uses information from adjacent branches.", "topics": ["data-generation", "planning", "memory", "efficiency"], "updated_at": "2026-02-10", "url": "https://arxiv.org/abs/2602.09662"}
{"aliases": [], "authors": ["Yuhao Zheng", "Li'an Zhong", "Yi Wang", "Rui Dai", "Kaikui Liu", "Xiangxiang Chu", "Linyuan Lv", "Philip Torr", "Kevin Qinghong Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "55287e2fe8a5e94671e4aab8e9d4a907ba9b6c83306d9b3f61b74adccd5c5059", "submitted_at": "2026-02-10T14:56:19Z", "updated_at": "2026-02-10T14:56:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.09856v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2602.09856", "identifiers": {"arxiv": "2602.09856"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09856"}], "name": "Code2World: A GUI World Model via Renderable Code Generation", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Code2World predicts future GUI screens through renderable code, trained on AndroidCode trajectories with supervised learning and render-aware reinforcement learning. It evaluates visual prediction and downstream AndroidWorld navigation with the learned world model.", "topics": ["world-models", "data-generation", "reinforcement-learning", "mobile-navigation"], "updated_at": "2026-02-10", "url": "https://arxiv.org/abs/2602.09856"}
{"aliases": [], "authors": ["Eranga Bandara", "Ross Gore", "Sachin Shetty", "Sachini Rajapakse", "Isurunima Kularathna", "Pramoda Karunarathna", "Ravi Mukkamala", "Peter Foytik", "Safdar H. Bouk", "Abdul Rahman", "Xueping Liang", "Amin Hass", "Tharaka Hewa", "Ng Wee Keong", "Kasun De Zoysa", "Aruna Withanage", "Nilaan Loganathan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5b858b7d99814d8b3bc8c926f7b5df6fb8f1f760747229f969bba08a9d63b64a", "submitted_at": "2026-01-27T10:49:59Z", "updated_at": "2026-01-27T10:49:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.10122v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.10122", "identifiers": {"arxiv": "2602.10122"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.10122"}], "name": "A Practical Guide to Agentic AI Transition in Organizations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-27", "url": "https://arxiv.org/abs/2602.10122"}
{"aliases": [], "authors": ["Lepeng Zhao", "Zhenhua Zou", "Shuo Li", "Zhuotao Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "727279bf83a67f66fa8f173d028936fe0bdae63e9f23800d98476da10201ad0d", "submitted_at": "2026-02-08T15:50:04Z", "updated_at": "2026-04-26T01:34:23Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.10139v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.10139", "identifiers": {"arxiv": "2602.10139"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.10139"}], "name": "Anonymization-Enhanced Privacy Protection for Mobile GUI Agents: Available but Invisible", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "An anonymization framework replaces sensitive screen content and instructions with consistent, type-preserving placeholders while mediating actions through a local interaction proxy. Experiments on AndroidLab and PrivScreen examine the privacy–utility trade-off when mobile GUI agents operate over these transformed interfaces.", "topics": ["safety-security", "screen-understanding", "mobile-navigation"], "updated_at": "2026-04-26", "url": "https://arxiv.org/abs/2602.10139"}
{"aliases": [], "authors": ["Tianci Xue", "Zeyi Liao", "Tianneng Shi", "Zilu Wang", "Kai Zhang", "Dawn Song", "Yu Su", "Huan Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3385966e7b483b33738b3c8e6c9a4d9cf17afa671e1f2e49de4f946cbc5f251", "submitted_at": "2026-02-10T23:06:02Z", "updated_at": "2026-05-11T00:11:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.10356v2"}}, "environments": ["desktop", "linux", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:29:29+00:00", "role": "primary", "sha256": "fba520c76ca9268af9239fbeaf5836e21932ab41ce53ccba71f47e3fc767028a", "supports": ["summary", "environments", "scope"], "url": "https://arxiv.org/html/2602.10356v2"}], "id": "arxiv:2602.10356", "identifiers": {"arxiv": "2602.10356"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.10356"}], "name": "Autonomous Continual Learning for Environment Adaptation of Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract and the specified sections of the same recorded paper version. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "ACuRL uses autonomous interface exploration and an adaptive task curriculum to continually train computer-use policies with reinforcement learning. CUAJudge supplies outcome feedback, and experiments study adaptation across desktop applications and environment changes.", "topics": ["reinforcement-learning", "data-generation", "reward-models", "robustness"], "updated_at": "2026-05-11", "url": "https://arxiv.org/abs/2602.10356"}
{"aliases": [], "authors": ["George Pu", "Michael S. Lee", "Udari Madhushani Sehwag", "David J. Lee", "Bryan Zhu", "Yash Maurya", "Mohit Raghavendra", "Yuan Xue", "Samuel Marc Denton"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15c13bf220b8178b66da9151e013afd08d6b61cf029e4feed7c13a74202dbf0f", "submitted_at": "2026-02-11T04:49:50Z", "updated_at": "2026-03-20T17:00:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.10525v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.10525", "identifiers": {"arxiv": "2602.10525"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.10525"}], "name": "LHAW: Controllable Underspecification for Long-Horizon Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-20", "url": "https://arxiv.org/abs/2602.10525"}
{"aliases": [], "authors": ["Xingyi Zhang", "Yulei Ye", "Kaifeng Huang", "Wenhao Li", "Xiangfeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86a47f084b23cc70e754808f6404bed0bbd500f78fcac4ccc104f95052a49868", "submitted_at": "2026-02-11T12:54:53Z", "updated_at": "2026-02-11T12:54:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.10814v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.10814", "identifiers": {"arxiv": "2602.10814"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.10814"}], "name": "See, Plan, Snap: Evaluating Multimodal GUI Agents in Scratch", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "ScratchWorld evaluates program construction through the Scratch interface using fine-grained drag-and-drop actions and a complementary semantic-API mode. Runtime tests assess the resulting programs, allowing analysis of the gap between program reasoning and GUI manipulation.", "topics": ["evaluation", "grounding", "automation"], "updated_at": "2026-02-11", "url": "https://arxiv.org/abs/2602.10814"}
{"aliases": [], "authors": ["Zhenhua Zou", "Sheng Guo", "Qiuyang Zhan", "Lepeng Zhao", "Shuo Li", "Qi Li", "Ke Xu", "Mingwei Xu", "Zhuotao Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f973754a1630fac3c37307f39fc7346601f2bf269a0a2804e8c0841895d1a9d5", "submitted_at": "2026-02-11T14:52:27Z", "updated_at": "2026-02-13T12:42:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.10915v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2602.10915", "identifiers": {"arxiv": "2602.10915"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.10915"}], "name": "Blind Gods and Broken Screens: Architecting a Secure, Intent-Centric Mobile Agent Operating System", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-13", "url": "https://arxiv.org/abs/2602.10915"}
{"aliases": [], "authors": ["Hubert M. Pysklo", "Artem Zhuravel", "Patrick D. Watson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "96ac36025b1fb4753426165f82ff41d290a1dd7cdd412b5454b58243519d80bd", "submitted_at": "2026-02-11T13:31:18Z", "updated_at": "2026-04-28T16:08:25Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.11224v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.11224", "identifiers": {"arxiv": "2602.11224"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.11224"}], "name": "Agent-Diff: Benchmarking LLM Agents on Enterprise API Tasks via Code Execution with State-Diff-Based Evaluation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-28", "url": "https://arxiv.org/abs/2602.11224"}
{"aliases": [], "authors": ["Sidong Feng", "Chunyang Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb8051355f7cff469b701b0ef0243fc1425c5fc74eb3894c33aee002f9abe8b1", "submitted_at": "2026-02-12T03:14:11Z", "updated_at": "2026-02-12T03:14:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.11514v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.11514", "identifiers": {"arxiv": "2602.11514"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.11514"}], "name": "How Smart Is Your GUI Agent? A Framework for the Future of Software Interaction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-12", "url": "https://arxiv.org/abs/2602.11514"}
{"aliases": [], "authors": ["Congmin Zheng", "Xiaoyun Mo", "Xinbei Ma", "Qiqiang Lin", "Yin Zhao", "Jiachen Zhu", "Xingyu Lou", "Jun Wang", "Zhaoxiang Wang", "Weiwen Liu", "Zhuosheng Zhang", "Yong Yu", "Weinan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ffec5a47e31b541cfdf0fb3c143b55cbb28119463731d9a6f3add0fe116caea9", "submitted_at": "2026-02-12T03:31:40Z", "updated_at": "2026-02-12T03:31:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.11524v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.11524", "identifiers": {"arxiv": "2602.11524"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.11524"}], "name": "Adaptive Milestone Reward for GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "ADMIRE extracts milestones from successful GUI explorations to provide adaptive intermediate rewards. It assigns credit differently to successful and failed trajectories and evaluates training on AndroidWorld, with additional transfer studies in web and embodied settings.", "topics": ["reinforcement-learning", "reward-models", "mobile-navigation"], "updated_at": "2026-02-12", "url": "https://arxiv.org/abs/2602.11524"}
{"aliases": [], "authors": ["Xiwen Teoh", "Yun Lin", "Duc-Minh Nguyen", "Ruofei Ren", "Wenjie Zhang", "Jin Song Dong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2076fc8b0b7b37bdf43fe2e0b0cc5fa4c3a8cb7b44a41fdf80f59a8c05c7e1b3", "submitted_at": "2026-02-12T08:51:07Z", "updated_at": "2026-04-23T15:00:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.11724v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2602.11724", "identifiers": {"arxiv": "2602.11724"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.11724"}], "name": "WebTestPilot: Agentic End-to-End Web Testing against Natural Language Specification by Inferring Oracles with Symbolized GUI Elements", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebTestPilot turns natural-language requirements into web test steps with inferred preconditions and postconditions over symbolized GUI elements. A benchmark of bug-injected web applications tests whether these context-aware oracles distinguish application defects from agent mistakes.", "topics": ["automation", "evaluation", "planning"], "updated_at": "2026-04-23", "url": "https://arxiv.org/abs/2602.11724"}
{"aliases": [], "authors": ["Jiazheng Sun", "Mingxuan Li", "Yingying Zhang", "Jiayang Niu", "Yachen Wu", "Ruihan Jin", "Shuyu Lei", "Pengrongrui Tan", "Zongyu Zhang", "Ruoyi Wang", "Jiachen Yang", "Boyu Yang", "Jiacheng Liu", "Xin Peng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bbba09ca845f4fb772516d26d59fc3d99f8e87ba3657bb29c3c062b50925ec5", "submitted_at": "2026-02-12T09:25:15Z", "updated_at": "2026-02-12T09:25:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.11750v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.11750", "identifiers": {"arxiv": "2602.11750"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.11750"}], "name": "AmbiBench: Benchmarking Mobile GUI Agents Beyond One-Shot Instructions in the Wild", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "AmbiBench evaluates mobile GUI agents under four levels of instruction clarity, including tasks requiring active clarification. Its MUSE evaluator separately audits outcomes, execution quality and interaction quality to study alignment with user intent.", "topics": ["evaluation", "human-in-the-loop", "mobile-navigation"], "updated_at": "2026-02-12", "url": "https://arxiv.org/abs/2602.11750"}
{"aliases": [], "authors": ["Lai Wei", "Liangbo He", "Jun Lan", "Lingzhong Dong", "Yutong Cai", "Siyuan Li", "Huijia Zhu", "Weiqiang Wang", "Linghe Kong", "Yue Wang", "Zhuosheng Zhang", "Weiran Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95ecf393e377979084c8fd3717fa2576f3d3fb48668b767c9c35832ab5473a85", "submitted_at": "2026-02-12T12:00:35Z", "updated_at": "2026-02-16T11:54:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.11858v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2602.11858", "identifiers": {"arxiv": "2602.11858"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.11858"}], "name": "Zooming without Zooming: Region-to-Image Distillation for Fine-Grained Multimodal Perception", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-16", "url": "https://arxiv.org/abs/2602.11858"}
{"aliases": [], "authors": ["Zhen Zhang", "Kaiqiang Song", "Xun Wang", "Yebowen Hu", "Weixiang Yan", "Chenyang Zhao", "Henry Peng Zou", "Haoyun Deng", "Sathish Reddy Indurthi", "Shujian Liu", "Simin Ma", "Xiaoyang Wang", "Xin Eric Wang", "Song Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2c7166f28c87f1d857162a3b734bb05152338f60a802efdb98506376f545d35", "submitted_at": "2026-02-12T18:55:09Z", "updated_at": "2026-02-20T21:00:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.12268v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.12268", "identifiers": {"arxiv": "2602.12268"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.12268"}], "name": "CM2: Reinforcement Learning with Checklist Rewards for Multi-Turn and Multi-Step Agentic Tool Use", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-20", "url": "https://arxiv.org/abs/2602.12268"}
{"aliases": [], "authors": ["Nicholas Lee", "Lutfi Eren Erdogan", "Chris Joseph John", "Surya Krishnapillai", "Michael W. Mahoney", "Kurt Keutzer", "Amir Gholami"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c64cd65fc7fec601658baf7836bdbe9c0e494bde2b0a4b66ec9d3090ee106b1f", "submitted_at": "2026-02-12T18:58:30Z", "updated_at": "2026-08-14T21:25:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.12276v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2602.12276", "identifiers": {"arxiv": "2602.12276"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.12276"}], "name": "Agentic Test-Time Scaling for WebAgents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "CATTS allocates additional inference computation when an agent’s action votes show uncertainty. The paper compares voting and arbiter strategies and evaluates this selective scaling on WebArena-Lite, Online-Mind2Web and GoBrowse.", "topics": ["efficiency", "planning", "evaluation", "web-navigation"], "updated_at": "2026-08-14", "url": "https://arxiv.org/abs/2602.12276"}
{"aliases": [], "authors": ["Renjun Xu", "Yang Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0edc70265a8d46a7c22a9c6384de8aa19f79dbb6987f8c23f4a260647a6a3f0", "submitted_at": "2026-02-12T21:33:25Z", "updated_at": "2026-06-02T16:14:54Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2602.12430v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.12430", "identifiers": {"arxiv": "2602.12430"}, "kind": "survey", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.12430"}], "name": "Agent Skills for Large Language Models: Architecture, Acquisition, Security, and the Path Forward", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent survey of the skill layer: CUA deployment and GUI grounding are explicit parts of the review, but its overall scope extends well beyond interface agents.", "published_at": "2026-02-12", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "This survey organizes agent skills by architecture, acquisition, deployment and security, including their role in computer-use systems and GUI grounding. It also proposes a permission and lifecycle governance framework for skill deployment.", "topics": ["survey", "tool-integration", "safety-security", "planning"], "updated_at": "2026-06-02", "url": "https://arxiv.org/abs/2602.12430"}
{"aliases": [], "authors": ["Lajanugen Logeswaran", "Jaekyeom Kim", "Sungryull Sohn", "Creighton Glasscock", "Honglak Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "555285a1ab51e02661067bdf8bcc3fa7f6ee0fdd63f52f786a03eb2483bd1a31", "submitted_at": "2026-02-13T02:52:18Z", "updated_at": "2026-02-13T02:52:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.12544v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.12544", "identifiers": {"arxiv": "2602.12544"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.12544"}], "name": "Scaling Web Agent Training through Automatic Data Generation and Fine-grained Evaluation", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "This work uses constraint-based progress evaluation to retain partially successful web trajectories for training. It introduces BookingArena for structured booking tasks and studies a distilled agent trained with the resulting data.", "topics": ["data-generation", "evaluation", "web-navigation"], "updated_at": "2026-02-13", "url": "https://arxiv.org/abs/2602.12544"}
{"aliases": [], "authors": ["Junjie Wang", "Zequn Xie", "Dan Yang", "Jie Feng", "Yue Shen", "Duolin Sun", "Meixiu Long", "Yihan Jiao", "Zhehao Tan", "Jian Wang", "Peng Wei", "Jinjie Gu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7ebb062fd2746028659aebccb5e445bb1755c4def5a6f14b3a8a2e45c8d6d7dc", "submitted_at": "2026-02-13T11:56:20Z", "updated_at": "2026-05-08T07:17:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.12852v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.12852", "identifiers": {"arxiv": "2602.12852"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.12852"}], "name": "WebClipper: Efficient Evolution of Web Agents with Graph-based Trajectory Pruning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2602.12852"}
{"aliases": [], "authors": ["Daesik Jang", "Morgan Lindsay Heisler", "Linzi Xing", "Yifei Li", "Edward Wang", "Ying Xiong", "Yong Zhang", "Zhenan Fan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db10440d042227394c93142f40b6c6d0bed2e9a0923af9c3c54359ce9f892b82", "submitted_at": "2026-02-10T19:49:06Z", "updated_at": "2026-02-10T19:49:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.13318v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.13318", "identifiers": {"arxiv": "2602.13318"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.13318"}], "name": "DECKBench: Benchmarking Multi-Agent Frameworks for Academic Slide Generation and Editing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-10", "url": "https://arxiv.org/abs/2602.13318"}
{"aliases": [], "authors": ["Jaechul Roh", "Eugene Bagdasarian", "Hamed Haddadi", "Ali Shahin Shamsabadi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb048b4e5d1d244d689957f30cdcb93cf4dbb750b24ae527b62ecfb04b554eb4", "submitted_at": "2026-02-13T23:02:50Z", "updated_at": "2026-02-13T23:02:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.13516v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2602.13516", "identifiers": {"arxiv": "2602.13516"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.13516"}], "name": "SPILLage: Agentic Oversharing on the Web", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Formalizes and measures natural agentic oversharing when browser agents handle private user emails and calendars across live websites.", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2026-02-13", "url": "https://arxiv.org/abs/2602.13516"}
{"aliases": [], "authors": ["Yunfan Zhang", "Kathleen McKeown", "Smaranda Muresan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "380b87055cdd9167a2ffcd121dc2ca848fe2c3c6ede34365007ef8c23177e7fb", "submitted_at": "2026-02-14T01:18:51Z", "updated_at": "2026-02-14T01:18:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.13543v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.13543", "identifiers": {"arxiv": "2602.13543"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.13543"}], "name": "LiveNewsBench: Evaluating LLM Web Search Capabilities with Freshly Curated News", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-14", "url": "https://arxiv.org/abs/2602.13543"}
{"aliases": [], "authors": ["Yuyu Guo", "Wenjie Yang", "Siyuan Yang", "Ziyang Liu", "Cheng Chen", "Yuan Wei", "Yun Hu", "Yang Huang", "Guoliang Hao", "Dongsheng Yuan", "Jianming Wang", "Xin Chen", "Hang Yu", "Lei Lei", "Peng Di"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "04298ed8a4bc0f2811f52eeee6da0a13faa265bb3c5cbc5ea6be573a77c21f27", "submitted_at": "2026-02-14T02:33:55Z", "updated_at": "2026-04-30T02:05:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.13559v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2602.13559", "identifiers": {"arxiv": "2602.13559"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.13559"}], "name": "OpAgent: Operator Agent for Web Navigation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-02-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "OpAgent combines multi-task GUI fine-tuning with online web reinforcement learning using outcome judgments and rule-based progress rewards. A modular planner, grounder, reflector and summarizer support navigation and recovery, evaluated separately from the underlying trained policy.", "topics": ["web-navigation", "reinforcement-learning", "planning", "self-reflection"], "updated_at": "2026-04-30", "url": "https://arxiv.org/abs/2602.13559"}
{"aliases": [], "authors": ["Yibo Wang", "Guangda Huzhang", "Yuwei Hu", "Yu Xia", "Shiyin Lu", "Qing-Guo Chen", "Zhao Xu", "Weihua Luo", "Kaifu Zhang", "Lijun Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "573d859133ec9986af92e1943d0ec913d914acf4e54a030a4b103181ba505ae0", "submitted_at": "2026-02-14T07:44:47Z", "updated_at": "2026-02-14T07:44:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.13653v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2602.13653", "identifiers": {"arxiv": "2602.13653"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.13653"}], "name": "Building Autonomous GUI Navigation via Agentic-Q Estimation and Step-Wise Policy Optimization", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "This framework learns an agentic-Q model to estimate individual actions’ contributions to GUI task completion. It then optimizes a policy with step-level samples from its own trajectories, decoupling policy updates from environment interaction.", "topics": ["reinforcement-learning", "reward-models", "grounding"], "updated_at": "2026-02-14", "url": "https://arxiv.org/abs/2602.13653"}
{"aliases": [], "authors": ["Hao Kang", "Ziyang Li", "Weili Xu", "Xinyu Yang", "Yinfang Chen", "Junxiong Wang", "Beidi Chen", "Tushar Krishna", "Chenfeng Xu", "Simran Arora"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cd9bf2c90aca0167e1f9882ddbb3fa6b019532df37b82f1e0f57b2b3f1ecb99b", "submitted_at": "2026-02-14T09:26:41Z", "updated_at": "2026-06-30T17:19:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.13692v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.13692", "identifiers": {"arxiv": "2602.13692"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.13692"}], "name": "ThunderAgent: A Simple, Fast and Program-Aware Agentic Inference System", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-30", "url": "https://arxiv.org/abs/2602.13692"}
{"aliases": [], "authors": ["Mingfei Lu", "Mengjia Wu", "Feng Liu", "Jiawei Xu", "Weikai Li", "Haoyang Wang", "Zhengdong Hu", "Ying Ding", "Yizhou Sun", "Jie Lu", "Yi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "36b5621e54979becee17a4fc02fbd6402387acf1c023780a1fcf7c733d095c2c", "submitted_at": "2026-02-15T07:56:24Z", "updated_at": "2026-02-15T07:56:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14038v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.14038", "identifiers": {"arxiv": "2602.14038"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.14038"}], "name": "Choosing How to Remember: Adaptive Memory Structures for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-15", "url": "https://arxiv.org/abs/2602.14038"}
{"aliases": [], "authors": ["Yuan Cao", "Dezhi Ran", "Mengzhou Wu", "Yuzhe Guo", "Xin Chen", "Ang Li", "Gang Cao", "Gong Zhi", "Hao Yu", "Linyi Li", "Wei Yang", "Tao Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29b4c8a5bda1df237fad33ce483018b3476684f904841ed0598ec4286234b858", "submitted_at": "2026-02-15T10:58:01Z", "updated_at": "2026-02-15T10:58:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14093v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.14093", "identifiers": {"arxiv": "2602.14093"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14093"}], "name": "GUI-GENESIS: Automated Synthesis of Efficient Environments with Verifiable Rewards for GUI Agent Post-Training", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-GENESIS reconstructs applications as lightweight web environments with executable assertions for reward computation. It evaluates the cost of these training environments and transfer from synthetic interaction training to held-out real tasks.", "topics": ["data-generation", "sandbox", "reinforcement-learning", "evaluation"], "updated_at": "2026-02-15", "url": "https://arxiv.org/abs/2602.14093"}
{"aliases": [], "authors": ["A. Said Gurbuz", "Sunghwan Hong", "Ahmed Nassar", "Marc Pollefeys", "Peter Staar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d7c6e23f704562c4ae4db9e0b4f8f2a8dde1318efc08876ee57f5967b178257e", "submitted_at": "2026-02-15T19:00:02Z", "updated_at": "2026-05-01T17:32:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.14276v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2602.14276", "identifiers": {"arxiv": "2602.14276"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14276"}], "name": "ScreenParse: Moving Beyond Sparse Grounding with Complete Screen Parsing Supervision", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "ScreenParse supplies dense annotations of UI elements, types and text in web screenshots through an automated rendering and filtering pipeline. ScreenVLM learns to output structured screen markup, and the study evaluates dense parsing and transfer to GUI grounding.", "topics": ["screen-understanding", "data-generation", "grounding", "evaluation"], "updated_at": "2026-05-01", "url": "https://arxiv.org/abs/2602.14276"}
{"aliases": [], "authors": ["Yifan Wu", "Yiran Peng", "Yiyu Chen", "Jianhao Ruan", "Zijie Zhuang", "Cheng Yang", "Jiayi Zhang", "Man Chen", "Yenchi Tseng", "Zhaoyang Yu", "Liang Chen", "Yuyao Zhai", "Bang Liu", "Chenglin Wu", "Yuyu Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3773ec51f5316b9006f93aeafd7b654bf0e7418253c0af16ef7b7c00fddea102", "submitted_at": "2026-02-15T20:03:19Z", "updated_at": "2026-02-15T20:03:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14296v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2602.14296", "identifiers": {"arxiv": "2602.14296"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14296"}], "name": "AutoWebWorld: Synthesizing Infinite Verifiable Web Environments via Finite State Machines", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.", "published_at": "2026-02-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "AutoWebWorld turns finite-state task models into interactive websites whose actions and goal states can be checked programmatically. It generates verified trajectories for training and evaluates transfer to WebVoyager and Online-Mind2Web.", "topics": ["data-generation", "sandbox", "web-navigation", "evaluation"], "updated_at": "2026-02-15", "url": "https://arxiv.org/abs/2602.14296"}
{"aliases": [], "authors": ["Yukang Feng", "Jianwen Sun", "Zelai Yang", "Jiaxin Ai", "Chuanhao Li", "Zizhen Li", "Fanrui Zhang", "Kang He", "Rui Ma", "Jifan Lin", "Jie Sun", "Yang Xiao", "Sizhuo Zhou", "Wenxiao Wu", "Yiming Liu", "Pengfei Liu", "Yu Qiao", "Shenglin Zhang", "Kaipeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54a58bfdd017c1bf1fdfb26f2ffe0c22e0f1f65ed0af0b605fbefc462d7350df", "submitted_at": "2026-02-15T23:12:57Z", "updated_at": "2026-02-26T11:54:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.14337v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.14337", "identifiers": {"arxiv": "2602.14337"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.14337"}], "name": "LongCLI-Bench: A Preliminary Benchmark and Study for Long-horizon Agentic Programming in Command-Line Interfaces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-26", "url": "https://arxiv.org/abs/2602.14337"}
{"aliases": [], "authors": ["Tianyu Chen", "Dongrui Liu", "Xia Hu", "Jingyi Yu", "Wenjie Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ec0af138b01a510eabe6a062c077b1ae4ee02cfd5ba46b8fa40cf1b77c3f2ca", "submitted_at": "2026-02-16T00:33:02Z", "updated_at": "2026-02-16T00:33:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14364v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.14364", "identifiers": {"arxiv": "2602.14364"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.14364"}], "name": "A Trajectory-Based Safety Audit of Clawdbot (OpenClaw)", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-16", "url": "https://arxiv.org/abs/2602.14364"}
{"aliases": [], "authors": ["Zikai Xiao", "Jianhong Tu", "Chuhang Zou", "Yuxin Zuo", "Zhi Li", "Peng Wang", "Bowen Yu", "Fei Huang", "Junyang Lin", "Zuozhu Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2dca0e4219a788205bbb060911d0f004d1db4cf2ff8a4962c04fd5777ea194d7", "submitted_at": "2026-02-16T13:06:49Z", "updated_at": "2026-02-16T13:06:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14721v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.14721", "identifiers": {"arxiv": "2602.14721"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14721"}], "name": "WebWorld: A Large-Scale World Model for Web Agent Training", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-02-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebWorld, an open-web simulator trained on interaction trajectories for generating agent training data and testing candidate actions. WebWorld-Bench evaluates simulated transitions, alongside downstream web-agent training experiments.", "topics": ["world-models", "data-generation", "evaluation"], "updated_at": "2026-02-16", "url": "https://arxiv.org/abs/2602.14721"}
{"aliases": [], "authors": ["Chenyang Ma", "Clyde Fare", "Matthew Wilson", "Dave Braines"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "353036db8a5609d64792ea49523ca424cf265d981ff34d1e7bc8fed35ffb21c0", "submitted_at": "2026-02-16T15:59:56Z", "updated_at": "2026-02-16T15:59:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14865v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.14865", "identifiers": {"arxiv": "2602.14865"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14865"}], "name": "EmbeWebAgent: Embedding Web Agents into Any Customized UI", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-02-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "EmbeWebAgent embeds agent access into existing web applications through frontend observations and a page-specific function registry. Its workflow combines GUI primitives, composite actions and MCP tools, demonstrated in a live interface.", "topics": ["web-navigation", "tool-integration", "automation", "workflow"], "updated_at": "2026-02-16", "url": "https://arxiv.org/abs/2602.14865"}
{"aliases": [], "authors": ["Zhouzhou Shen", "Xueyu Hu", "Xiyun Li", "Tianqing Fang", "Juncheng Li", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29fa4927311278ad8ad8aabc8b77dfc05ad7d21b87969b5fb330d7a3a8fd727b", "submitted_at": "2026-02-17T06:37:31Z", "updated_at": "2026-02-17T06:37:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.15384v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.15384", "identifiers": {"arxiv": "2602.15384"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.15384"}], "name": "World-Model-Augmented Web Agents with Action Correction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-17", "url": "https://arxiv.org/abs/2602.15384"}
{"aliases": [], "authors": ["Zexue He", "Yu Wang", "Churan Zhi", "Yuanzhe Hu", "Tzu-Ping Chen", "Lang Yin", "Ze Chen", "Tong Arthur Wu", "Siru Ouyang", "Zihan Wang", "Jiaxin Pei", "Julian McAuley", "Yejin Choi", "Alex Pentland"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2317a94d54b4458a1b7c5e92754aeaefcbf270aabfa0bd8ea333c673cd4596b3", "submitted_at": "2026-02-18T09:49:14Z", "updated_at": "2026-09-17T22:55:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.16313v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.16313", "identifiers": {"arxiv": "2602.16313"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.16313"}], "name": "MemoryArena: Benchmarking Agent Memory in Interdependent Multi-Session Agentic Tasks", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-02-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "MemoryArena links tasks across sessions so that agents must distill earlier interaction experience and use it in later decisions. Its evaluation includes web navigation alongside planning, search and formal reasoning, testing memory through subsequent actions rather than recall alone.", "topics": ["memory", "interaction-history", "planning", "evaluation", "long-horizon"], "updated_at": "2026-09-17", "url": "https://arxiv.org/abs/2602.16313"}
{"aliases": [], "authors": ["Maksim Elistratov", "Marina Barannikov", "Gregory Ivanov", "Valentin Khrulkov", "Anton Konushin", "Andrey Kuznetsov", "Dmitrii Zhemchuzhnikov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4be653b49cc62353d6b3ecb3690a780131496037aef7ec067ba3f367aba2e48", "submitted_at": "2026-02-18T09:54:57Z", "updated_at": "2026-02-18T09:54:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.16317v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.16317", "identifiers": {"arxiv": "2602.16317"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.16317"}], "name": "CADEvolve: Creating Realistic CAD via Program Evolution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-18", "url": "https://arxiv.org/abs/2602.16317"}
{"aliases": [], "authors": ["Stephan Rabanser", "Sayash Kapoor", "Peter Kirgis", "Kangheng Liu", "Saiteja Utpala", "Arvind Narayanan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e2c0c133a0a57157b80e12b27d03000cd43a667ec8f74d0c04520ae30f2043ab", "submitted_at": "2026-02-18T18:05:44Z", "updated_at": "2026-06-02T02:15:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.16666v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.16666", "identifiers": {"arxiv": "2602.16666"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.16666"}], "name": "Towards a Science of AI Agent Reliability", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-02", "url": "https://arxiv.org/abs/2602.16666"}
{"aliases": [], "authors": ["Haiyang Xu", "Xi Zhang", "Haowei Liu", "Junyang Wang", "Zhaozai Zhu", "Shengjie Zhou", "Xuhao Hu", "Feiyu Gao", "Junjie Cao", "Zihua Wang", "Zhiyuan Chen", "Jitong Liao", "Qi Zheng", "Jiahui Zeng", "Ze Xu", "Shuai Bai", "Junyang Lin", "Jingren Zhou", "Ming Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eeaddd8a32bdd4b53356ca43e849ea3697ac3d40a05b32e6c930dc445c243898", "submitted_at": "2026-02-15T01:52:19Z", "updated_at": "2026-02-15T01:52:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.16855v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.16855", "identifiers": {"arxiv": "2602.16855"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.16855"}], "name": "Mobile-Agent-v3.5: Multi-platform Fundamental GUI Agents", "notes": "Summary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.", "published_at": "2026-02-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents GUI-Owl-1.5 variants and a hybrid data pipeline for multimodal GUI interaction, tool use, memory, and multi-platform tasks.", "topics": ["tool-integration", "memory", "reinforcement-learning"], "updated_at": "2026-02-15", "url": "https://arxiv.org/abs/2602.16855"}
{"aliases": [], "authors": ["Tanqiu Jiang", "Yuhui Wang", "Jiacheng Liang", "Ting Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7ce1d157e148236ee764561e1e0c98f47aeaf6ec85bd05aeabf0081cf6555135", "submitted_at": "2026-02-18T21:30:20Z", "updated_at": "2026-02-18T21:30:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.16901v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.16901", "identifiers": {"arxiv": "2602.16901"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.16901"}], "name": "AgentLAB: Benchmarking LLM Agents against Long-Horizon Attacks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-18", "url": "https://arxiv.org/abs/2602.16901"}
{"aliases": [], "authors": ["Farnaz Zamiri Zeraati", "Yang Trista Cao", "Yuehan Qiao", "Hal Daumé", "Hernisa Kacorri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "85732f88b014b6fafca19ea20ec721f9256800874d0d822860f12f310dd0ca86", "submitted_at": "2026-02-18T22:44:43Z", "updated_at": "2026-02-18T22:44:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.16930v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.16930", "identifiers": {"arxiv": "2602.16930"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.16930"}], "name": "Say It My Way: Exploring Control in Conversational Visual Question Answering with Blind Users", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-18", "url": "https://arxiv.org/abs/2602.16930"}
{"aliases": [], "authors": ["Serin Kim", "Sangam Lee", "Dongha Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "28a23bf22920f8c33317e9d16f59d7085f188e79157301ff573abe773cf562ce", "submitted_at": "2026-02-19T01:54:26Z", "updated_at": "2026-05-27T12:15:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.17003v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.17003", "identifiers": {"arxiv": "2602.17003"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.17003"}], "name": "Persona2Web: Benchmarking Personalized Web Agents for Contextual Reasoning with User History", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Persona2Web tests whether web agents can resolve ambiguous requests using preferences implicit in a user’s history. Its reasoning-aware evaluation compares agent architectures and history-access strategies on personalized tasks on the open web.", "topics": ["evaluation", "memory", "web-navigation"], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2602.17003"}
{"aliases": [], "authors": ["Seoyoung Lee", "Seobin Yoon", "Seongbeen Lee", "Yoojung Chun", "Dayoung Park", "Doyeon Kim", "Joo Yong Sim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "32c3c5005eddd3f748bdfbfdb9d82fefc77cf11e77fcc36d74aae648df8f3103", "submitted_at": "2026-02-19T03:42:15Z", "updated_at": "2026-02-27T03:23:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.17049v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.17049", "identifiers": {"arxiv": "2602.17049"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.17049"}], "name": "IntentCUA: Learning Intent-level Representations for Skill Abstraction and Multi-Agent Planning in Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-27", "url": "https://arxiv.org/abs/2602.17049"}
{"aliases": [], "authors": ["Linxi Jiang", "Rui Xi", "Zhijie Liu", "Shuo Chen", "Zhiqiang Lin", "Suman Nath"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81cb12b4b7379a2f602ab9edeea4d1c3a2f16be486f93742829e7f65f2128968", "submitted_at": "2026-02-19T10:50:52Z", "updated_at": "2026-06-07T07:49:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.17245v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.17245", "identifiers": {"arxiv": "2602.17245"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.17245"}], "name": "Web Agents Should Use Typed Actions Instead of Click-Based Browsing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-07", "url": "https://arxiv.org/abs/2602.17245"}
{"aliases": [], "authors": ["Yiming Guan", "Rui Yu", "John Zhang", "Lu Wang", "Chaoyun Zhang", "Liqun Li", "Bo Qiao", "Si Qin", "He Huang", "Fangkai Yang", "Pu Zhao", "Lukas Wutschitz", "Samuel Kessler", "Huseyin A Inan", "Robert Sim", "Saravan Rajmohan", "Qingwei Lin", "Dongmei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d2237dd43f4c876ba69b68c1871ca3efd3aa501e93c86655eb2de6d9f35d2e8", "submitted_at": "2026-02-19T13:48:29Z", "updated_at": "2026-02-19T13:48:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.17365v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.17365", "identifiers": {"arxiv": "2602.17365"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.17365"}], "name": "Computer-Using World Model", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-02-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents a desktop world model that predicts textual UI changes before synthesizing the next screenshot. It learns from Microsoft Office transitions and is evaluated by letting a frozen agent compare simulated actions before execution.", "topics": ["world-models", "planning", "action-prediction"], "updated_at": "2026-02-19", "url": "https://arxiv.org/abs/2602.17365"}
{"aliases": [], "authors": ["Yue Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4e3f4fb337b375bdb84456ed3128b1b282d7c97627795e9f3b09534dc784dc57", "submitted_at": "2026-02-19T17:01:08Z", "updated_at": "2026-04-25T16:40:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.17547v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.17547", "identifiers": {"arxiv": "2602.17547"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.17547"}], "name": "KLong: Training LLM Agent for Extremely Long-horizon Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-25", "url": "https://arxiv.org/abs/2602.17547"}
{"aliases": [], "authors": ["Faria Huq", "Zora Zhiruo Wang", "Zhanqiu Guo", "Venu Arvind Arangarajan", "Tianyue Ou", "Frank Xu", "Shuyan Zhou", "Graham Neubig", "Jeffrey P. Bigham"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe83d38ece7cd8f39d932611abaf7a170e7cb9b7c9744bd1176eb42e73be85b5", "submitted_at": "2026-02-19T18:11:28Z", "updated_at": "2026-07-07T23:37:46Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2602.17588v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.17588", "identifiers": {"arxiv": "2602.17588"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.17588"}], "name": "Modeling Distinct Human Interaction in Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "CowCorpus records interleaved human and agent actions during web navigation and identifies distinct patterns of user intervention. Models trained to anticipate these interventions are integrated into live web agents and assessed in a user study.", "topics": ["human-in-the-loop", "web-navigation", "data-generation", "action-prediction"], "updated_at": "2026-07-07", "url": "https://arxiv.org/abs/2602.17588"}
{"aliases": [], "authors": ["Sima Noorani", "Shayan Kiyani", "Hamed Hassani", "George Pappas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7ca65b664159c062c6ba0f06f542f436257984e4e5a4c01cf160213a52c00143", "submitted_at": "2026-02-19T18:54:34Z", "updated_at": "2026-02-24T18:15:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.17646v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.17646", "identifiers": {"arxiv": "2602.17646"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.17646"}], "name": "Multi-Round Human-AI Collaboration with User-Specified Requirements", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-24", "url": "https://arxiv.org/abs/2602.17646"}
{"aliases": [], "authors": ["Aditya Kumar Singh", "Hitesh Kandala", "Pratik Prabhanjan Brahma", "Zicheng Liu", "Emad Barsoum"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "74db7c0382928a2932cdc7c75f4dcfef780d70adc400996e1df318646e292c86", "submitted_at": "2026-02-21T14:22:49Z", "updated_at": "2026-03-27T15:03:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.18846v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.18846", "identifiers": {"arxiv": "2602.18846"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.18846"}], "name": "DUET-VLM: Dual stage Unified Efficient Token reduction for VLM Training and Inference", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-27", "url": "https://arxiv.org/abs/2602.18846"}
{"aliases": [], "authors": ["Dongming Jiang", "Yi Li", "Songtao Wei", "Jinxin Yang", "Ayushi Kishore", "Alysa Zhao", "Dingyi Kang", "Xu Hu", "Feng Chen", "Qiannan Li", "Bingzhe Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "869a21cc40665fb6b790a236506bad080879472ebc4f5f07b55fa7761f621987", "submitted_at": "2026-02-22T19:50:01Z", "updated_at": "2026-05-20T06:06:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.19320v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.19320", "identifiers": {"arxiv": "2602.19320"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.19320"}], "name": "Anatomy of Agentic Memory: Taxonomy and Empirical Analysis of Evaluation and System Limitations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-20", "url": "https://arxiv.org/abs/2602.19320"}
{"aliases": [], "authors": ["Ayush Nangia", "Shikhar Mishra", "Aman Gokrani", "Paras Chopra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d493212d77426c1d674437efcef2366e839fde60803fd31cb13522b5d02571b2", "submitted_at": "2026-02-23T08:37:53Z", "updated_at": "2026-02-23T08:37:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.19594v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "id": "arxiv:2602.19594", "identifiers": {"arxiv": "2602.19594"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.19594"}], "name": "ISO-Bench: Can Coding Agents Optimize Real-World Inference Workloads?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-23", "url": "https://arxiv.org/abs/2602.19594"}
{"aliases": [], "authors": ["Qiankun Ma", "Ziyao Zhang", "Haofei Wang", "Jie Chen", "Zhen Song", "Hairong Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aaa63b3d7c3cb76acb2cef4ab5c29497c3c170f653703d1b92374f10e05f1b6b", "submitted_at": "2026-02-23T14:15:37Z", "updated_at": "2026-02-23T14:15:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.19870v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2602.19870", "identifiers": {"arxiv": "2602.19870"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.19870"}], "name": "ApET: Approximation-Error Guided Token Compression for Efficient VLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-23", "url": "https://arxiv.org/abs/2602.19870"}
{"aliases": [], "authors": ["Hongbin Zhong", "Fazle Faisal", "Luis França", "Tanakorn Leesatapornwongsa", "Adriana Szekeres", "Kexin Rong", "Suman Nath"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e179c833d43a88098d576e31e726eb3e4289f728659697364d0b448282df3fc1", "submitted_at": "2026-02-24T03:03:18Z", "updated_at": "2026-09-28T17:40:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.20502v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.20502", "identifiers": {"arxiv": "2602.20502"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.20502"}], "name": "ActionEngine: From Reactive to Programmatic Web Agents via State Machine Memory", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-02-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "ActionEngine explores web applications to build an updateable state-machine memory of GUI states, operations and transitions. It generates complete task programs from that memory, using reactive repair to handle changed interfaces or missing knowledge.", "topics": ["memory", "planning", "efficiency", "web-navigation"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2602.20502"}
{"aliases": [], "authors": ["Dezhi Kong", "Zhengzhao Feng", "Qiliang Liang", "Hao Wang", "Haofei Sun", "Changpeng Yang", "Yang Li", "Peng Zhou", "Shuai Nie", "Hongzhen Wang", "Linfeng Zhou", "Hao Jia", "Jiaming Xu", "Runyu Shi", "Ying Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0d668269627396b3c31bd4065e5658528702fb438e71abedecd31cfbccb635b", "submitted_at": "2026-02-25T12:32:37Z", "updated_at": "2026-05-08T09:18:26Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2602.21858v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2602.21858", "identifiers": {"arxiv": "2602.21858"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.21858"}], "name": "ProactiveMobile: A Comprehensive Benchmark for Boosting Proactive Intelligence on Mobile Devices", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2602.21858"}
{"aliases": [], "authors": ["Rui Yang", "Qianhui Wu", "Zhaoyang Wang", "Hanyang Chen", "Ke Yang", "Hao Cheng", "Huaxiu Yao", "Baolin Peng", "Huan Zhang", "Jianfeng Gao", "Tong Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "66298bdcd085f8921ad43cfdc9c2d4af7cdfd3ce5c900d77b49701956343b314", "submitted_at": "2026-02-25T18:34:57Z", "updated_at": "2026-05-25T21:32:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.22190v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.22190", "identifiers": {"arxiv": "2602.22190"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.22190"}], "name": "GUI-Libra: Training Native GUI Agents to Reason and Act with Action-aware Supervision and Partially Verifiable RL", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-25", "url": "https://arxiv.org/abs/2602.22190"}
{"aliases": [], "authors": ["Yujie Zhao", "Boqin Yuan", "Junbo Huang", "Haocheng Yuan", "Zhongming Yu", "Haozhou Xu", "Lanxiang Hu", "Abhilash Shankarampeta", "Zimeng Huang", "Wentao Ni", "Yuandong Tian", "Jishen Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ed187aca63afe99c22b98af8d078cd6300d194bf4b0382eda11d896b30283cfb", "submitted_at": "2026-02-26T08:59:31Z", "updated_at": "2026-05-27T09:28:43Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2602.22769v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2602.22769", "identifiers": {"arxiv": "2602.22769"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.22769"}], "name": "AMA-Bench: Evaluating Long-Horizon Memory for Agentic Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2602.22769"}
{"aliases": [], "authors": ["Zhaochen Su", "Jincheng Gao", "Hangyu Guo", "Zhenhua Liu", "Lueyang Zhang", "Xinyu Geng", "Shijue Huang", "Peng Xia", "Guanyu Jiang", "Cheng Wang", "Yue Zhang", "Yi R. Fung", "Junxian He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9051a7e95fbb8b024bff4b44d25cb9eb8aec2dfc8a75c2fe812375d0a4c25044", "submitted_at": "2026-02-26T16:30:46Z", "updated_at": "2026-03-02T03:29:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.23166v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2602.23166", "identifiers": {"arxiv": "2602.23166"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.23166"}], "name": "AgentVista: Evaluating Multimodal Agents in Ultra-Challenging Realistic Visual Scenarios", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-02", "url": "https://arxiv.org/abs/2602.23166"}
{"aliases": [], "authors": ["Zhou Xu", "Bowen Zhou", "Qi Wang", "Shuwen Feng", "Jingyu Xiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ac076fb0916a0ce436362f83d2b2dbece05e4abb8a854351c6a135f917e4bdc", "submitted_at": "2026-02-26T17:12:40Z", "updated_at": "2026-02-26T17:12:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.23235v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2602.23235", "identifiers": {"arxiv": "2602.23235"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.23235"}], "name": "Spatio-Temporal Token Pruning for Efficient High-Resolution GUI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-26", "url": "https://arxiv.org/abs/2602.23235"}
{"aliases": [], "authors": ["Tianjun Yao", "Yongqiang Chen", "Yujia Zheng", "Pan Li", "Zhiqiang Shen", "Kun Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e5b50ba5a79548fd35fc7a70abd79b15df8a88bc3700e86aef41077fa99ac62", "submitted_at": "2026-02-26T18:28:04Z", "updated_at": "2026-02-27T08:21:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.23320v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2602.23320", "identifiers": {"arxiv": "2602.23320"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.23320"}], "name": "ParamMem: Augmenting Language Agents with Parametric Reflective Memory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-27", "url": "https://arxiv.org/abs/2602.23320"}
{"aliases": [], "authors": ["Yuxuan Liu", "Weikai Xu", "Kun Huang", "Changyu Chen", "Jiankun Zhao", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Shuo Shang", "Bo Du", "Ji-Rong Wen", "Rui Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "46bc8b5d5097f41a29d37b3fdf6245778acd82f3ae800d5fec38b2c4f7d3aecb", "submitted_at": "2026-02-27T16:19:45Z", "updated_at": "2026-03-06T18:07:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.24142v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2602.24142", "identifiers": {"arxiv": "2602.24142"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2602.24142"}], "name": "CoME: Empowering Channel-of-Mobile-Experts with Informative Hybrid-Capabilities Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-06", "url": "https://arxiv.org/abs/2602.24142"}
{"aliases": [], "authors": ["Bowen Zhou", "Zhou Xu", "Wanli Li", "Jingyu Xiao", "Pingan Gan", "Haoqian Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03ac3b2fb60d64115d49e35ba867179c9f6a0f10a55f310b8b36cb4df82c6d53", "submitted_at": "2026-02-27T01:27:20Z", "updated_at": "2026-08-26T08:42:41Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.00188v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.00188", "identifiers": {"arxiv": "2603.00188"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.00188"}], "name": "ST-Lite: Training-Free KV Cache Compression with Spatio-Trajectory Guidance for Long-Horizon GUI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-26", "url": "https://arxiv.org/abs/2603.00188"}
{"aliases": [], "authors": ["Jindong Jiang", "Amala Sanjay Deshmukh", "Kateryna Chumachenko", "Karan Sapra", "Zhiding Yu", "Guilin Liu", "Andrew Tao", "Pavlo Molchanov", "Jan Kautz", "Wonmin Byeon"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a41f3e23f38b2963a77105a85ecda7ff1fc1539cd424fd7a1e10350d22866f7d", "submitted_at": "2026-02-27T08:11:06Z", "updated_at": "2026-07-01T00:40:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.00198v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.00198", "identifiers": {"arxiv": "2603.00198"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.00198"}], "name": "Stateful Token Reduction for Long-Video Hybrid VLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-01", "url": "https://arxiv.org/abs/2603.00198"}
{"aliases": [], "authors": ["Linxi Jiang", "Zhijie Liu", "Haotian Luo", "Zhiqiang Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3b24ac636d86124a70b44e84657b963db4c5af3367ae70e5c35e0bebb3f600f", "submitted_at": "2026-02-28T05:25:03Z", "updated_at": "2026-02-28T05:25:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.00476v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2603.00476", "identifiers": {"arxiv": "2603.00476"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.00476"}], "name": "Atomicity for Agents: Exposing, Exploiting, and Mitigating TOCTOU Vulnerabilities in Browser-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Identifies time-of-check to time-of-use (TOCTOU) race vulnerabilities when web pages mutate between agent planning and action execution.", "topics": ["safety-security", "web-navigation", "state-tracking"], "updated_at": "2026-02-28", "url": "https://arxiv.org/abs/2603.00476"}
{"aliases": [], "authors": ["Dawei Yan", "Haokui Zhang", "Guangda Huzhang", "Yang Li", "Yibo Wang", "Qing-Guo Chen", "Zhao Xu", "Weihua Luo", "Ying Li", "Wei Dong", "Chunhua Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1df893d9bb87b4d0a202c32c6c1b58dd4660f98d5406ed6bb0a5e5591890b931", "submitted_at": "2026-02-28T06:59:51Z", "updated_at": "2026-02-28T06:59:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.00503v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.00503", "identifiers": {"arxiv": "2603.00503"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.00503"}], "name": "M$^2$: Dual-Memory Augmentation for Long-Horizon Web Agents via Trajectory Summarization and Insight Retrieval", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-28", "url": "https://arxiv.org/abs/2603.00503"}
{"aliases": [], "authors": ["Zhe Wu", "Donglin Mo", "Hongjin Lu", "Junliang Xing", "Jianheng Liu", "Yuheng Jing", "Kai Li", "Kun Shao", "Jianye Hao", "Yuanchun Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4196cc95428b3bc00393b053bc5d12b72bc2b16440aa117e98fa7ef05d68f0cc", "submitted_at": "2026-02-28T14:33:14Z", "updated_at": "2026-02-28T14:33:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.00676v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2603.00676", "identifiers": {"arxiv": "2603.00676"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.00676"}], "name": "K^2-Agent: Co-Evolving Know-What and Know-How for Hierarchical Mobile Device Control", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-28", "url": "https://arxiv.org/abs/2603.00676"}
{"aliases": [], "authors": ["Kai Zhang", "Zhengqing Yuan", "Cheng Peng", "Songlin Zhao", "Mengxian Lyu", "Ziyi Chen", "Yanfang Ye", "Wei Liu", "Ying Zhang", "Kaleb E Smith", "Lifang He", "Lichao Sun", "Yonghui Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5ce55ab7086db74fd602552a8371ad68fe542bd5006243dbc347560e71fdb181", "submitted_at": "2026-03-01T00:06:43Z", "updated_at": "2026-09-21T07:31:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.00842v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.00842", "identifiers": {"arxiv": "2603.00842"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.00842"}], "name": "MedGPT-oss: Training a General-Purpose Vision-Language Model for Biomedicine", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-21", "url": "https://arxiv.org/abs/2603.00842"}
{"aliases": [], "authors": ["Martin Odersky", "Yaoyu Zhao", "Yichen Xu", "Oliver Bračevac", "Cao Nguyen Pham"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc90a7e8ace4608cb0165b2f81ed35bf82c84faa280f5d7b31c9618d95b676b0", "submitted_at": "2026-03-01T08:39:37Z", "updated_at": "2026-05-07T14:09:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.00991v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.00991", "identifiers": {"arxiv": "2603.00991"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.00991"}], "name": "Tracking Capabilities for Safer Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-07", "url": "https://arxiv.org/abs/2603.00991"}
{"aliases": [], "authors": ["Huanjin Yao", "Qixiang Yin", "Min Yang", "Ziwang Zhao", "Yibo Wang", "Haotian Luo", "Jingyi Zhang", "Jiaxing Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5d32a3465f5a0063358426cf6fd5434241d6191d225740f8e4f41466fd5e5fb8", "submitted_at": "2026-03-01T11:13:22Z", "updated_at": "2026-03-01T11:13:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.01050v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.01050", "identifiers": {"arxiv": "2603.01050"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.01050"}], "name": "MM-DeepResearch: A Simple and Effective Multimodal Agentic Search Baseline", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-01", "url": "https://arxiv.org/abs/2603.01050"}
{"aliases": [], "authors": ["Yutao Yang", "Junsong Li", "Qianjun Pan", "Bihao Zhan", "Yuxuan Cai", "Lin Du", "Jie Zhou", "Kai Chen", "Qin Chen", "Xin Li", "Bo Zhang", "Liang He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7e0aa13703bedc3673d64a61427a732e071cb54aacf8f06a5cc87d303fa77d0", "submitted_at": "2026-03-01T15:11:59Z", "updated_at": "2026-03-05T03:24:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.01145v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.01145", "identifiers": {"arxiv": "2603.01145"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.01145"}], "name": "AutoSkill: Experience-Driven Lifelong Learning via Skill Self-Evolution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-05", "url": "https://arxiv.org/abs/2603.01145"}
{"aliases": [], "authors": ["Tongzhou Wu", "Yuhao Wang", "Xinyu Ma", "Xiuqiang He", "Shuaiqiang Wang", "Dawei Yin", "Xiangyu Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fd295b5dfc35315af8fec5f900b06952bb762e04738453fd8c33fa96a01c956", "submitted_at": "2026-03-01T15:36:10Z", "updated_at": "2026-06-20T09:26:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.01152v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2603.01152", "identifiers": {"arxiv": "2603.01152"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.01152"}], "name": "DeepResearch-9K: A Challenging Benchmark Dataset of Deep-Research Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-20", "url": "https://arxiv.org/abs/2603.01152"}
{"aliases": [], "authors": ["Jiace Zhu", "Wentao Chen", "Qi Fan", "Zhixing Ren", "Junying Wu", "Xing Zhe Chai", "Chotiwit Rungrueangwutthinon", "Yehan Ma", "An Zou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e90c4c2cae8f5d603a45d8fa5c2693b49cd0d362694635a826c8c79ff0d1a1dc", "submitted_at": "2026-02-13T09:46:45Z", "updated_at": "2026-02-13T09:46:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.02236v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.02236", "identifiers": {"arxiv": "2603.02236"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.02236"}], "name": "CUDABench: Benchmarking LLMs for Text-to-CUDA Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-13", "url": "https://arxiv.org/abs/2603.02236"}
{"aliases": [], "authors": ["Ruiyang Zhang", "Qianguo Sun", "Chao Song", "Yiyan Qi", "Zhedong Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7a8734e35e0020f78fb7847add1a1caf7047b3e246f7d1eb0112ef8a57bd8524", "submitted_at": "2026-03-03T09:33:22Z", "updated_at": "2026-03-06T09:02:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.02795v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.02795", "identifiers": {"arxiv": "2603.02795"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.02795"}], "name": "VSearcher: Long-Horizon Multimodal Search Agent via Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-06", "url": "https://arxiv.org/abs/2603.02795"}
{"aliases": [], "authors": ["Zhenquan Yao", "Zitong Huang", "Yihan Zeng", "Jianhua Han", "Hang Xu", "Chun-Mei Feng", "Jianwei Ma", "Wangmeng Zuo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75f8eb625e969c2b180eeb3e90aa4e513ab5a1dc15657d3849c68b69bc5edf7c", "submitted_at": "2026-03-03T13:02:20Z", "updated_at": "2026-03-07T10:11:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.02951v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2603.02951", "identifiers": {"arxiv": "2603.02951"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.02951"}], "name": "CGL: Advancing Continual GUI Learning via Reinforcement Fine-Tuning", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The abstract describes artifact release as future work; this annotation does not assert current availability.", "published_at": "2026-03-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines supervised and reinforcement fine-tuning for continual GUI learning, using policy entropy to adjust their proportions and gradient projection to limit conflicting updates. AndroidControl-CL groups applications into sequential tasks to evaluate adaptation and retention.", "topics": ["reinforcement-learning", "robustness", "evaluation"], "updated_at": "2026-03-07", "url": "https://arxiv.org/abs/2603.02951"}
{"aliases": [], "authors": ["William Liang", "Sam Wang", "Hung-Ju Wang", "Osbert Bastani", "Yecheng Jason Ma", "Dinesh Jayaraman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4476650ab9ae181819a3e8876e155e1699e6f5f2c60ae6b022ba15f41409921a", "submitted_at": "2026-03-03T18:59:07Z", "updated_at": "2026-03-03T18:59:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.03278v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.03278", "identifiers": {"arxiv": "2603.03278"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.03278"}], "name": "Tether: Autonomous Functional Play with Correspondence-Driven Trajectory Warping", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-03", "url": "https://arxiv.org/abs/2603.03278"}
{"aliases": [], "authors": ["Haoyu Liu", "Dingcheng Li", "Lukas Rutishauser", "Zeyu Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "971770e61bf0af91ad1f17b197f09fc166b26bfb4953d3f2c74d457a87e25f3a", "submitted_at": "2026-03-04T18:29:54Z", "updated_at": "2026-10-05T23:01:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.04364v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.04364", "identifiers": {"arxiv": "2603.04364"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.04364"}], "name": "Dual-Modality Multi-Stage Adversarial Safety Training: Robustifying Multimodal Web Agents Against Cross-Modal Attacks", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies webpage injections that affect both screenshots and accessibility trees, then trains web agents through teacher imitation, supervised adversarial examples, and adversarial reinforcement learning. The evaluation tests defenses on MiniWoB++ and held-out tasks.", "topics": ["safety-security", "reinforcement-learning", "evaluation"], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2603.04364"}
{"aliases": [], "authors": ["Guilin Zhang", "Wei Jiang", "Xiejiashan Wang", "Aisha Behr", "Kai Zhao", "Jeffrey Friedman", "Xu Chu", "Amine Anoun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5d20c487b6502b34671a7e3a32006c19be81124379557bded5267f520e8cbada", "submitted_at": "2026-03-04T19:32:02Z", "updated_at": "2026-03-04T19:32:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.04549v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.04549", "identifiers": {"arxiv": "2603.04549"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.04549"}], "name": "Adaptive Memory Admission Control for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-04", "url": "https://arxiv.org/abs/2603.04549"}
{"aliases": [], "authors": ["Md Farhan Ishmam", "Kenneth Marino"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1901125b21431e41eacb5d5f0fc7a439d9a637c43762d441ece46a90e135051b", "submitted_at": "2026-03-05T08:43:06Z", "updated_at": "2026-10-14T10:16:21Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2603.04949v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.04949", "identifiers": {"arxiv": "2603.04949"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.04949"}], "name": "TimeWarp: Evaluating Web Agents by Revisiting the Past", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "TimeWarp tests web agents across multiple versions of the same sites, varying visual design, frontend code and workflows. TimeTraj transfers plans across versions to collect training trajectories and study robustness to interface change.", "topics": ["evaluation", "robustness", "web-navigation", "data-generation"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2603.04949"}
{"aliases": [], "authors": ["Sicheng Fan", "Qingyun Shi", "Shengze Xu", "Shengbo Cai", "Tieyong Zeng", "Li Ling", "Yanyi Shang", "Dehan Kong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fcca36d9a28ace835ede3fa5a93cb9c54608f164240c18bfe007bc945273b460", "submitted_at": "2026-03-05T10:51:34Z", "updated_at": "2026-04-14T08:14:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.05044v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.05044", "identifiers": {"arxiv": "2603.05044"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.05044"}], "name": "WebFactory: Automated Compression of Foundational Language Intelligence into Grounded Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebFactory, a GUI-agent training pipeline that synthesizes websites, generates tasks and trajectories, and trains with decomposed rewards. Offline and online transfer evaluations examine agents trained on the generated environments.", "topics": ["data-generation", "reinforcement-learning", "web-navigation"], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2603.05044"}
{"aliases": [], "authors": ["Sicheng Fan", "Rui Wan", "Yifei Leng", "Gaoning Liang", "Li Ling", "Yanyi Shang", "Dehan Kong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db52246d9f5f305046e6dd14f2ef15440c371678e1db0d05282c84a3667c6f10", "submitted_at": "2026-03-05T15:37:34Z", "updated_at": "2026-04-14T08:41:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.05295v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "id": "arxiv:2603.05295", "identifiers": {"arxiv": "2603.05295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.05295"}], "name": "WebChain: A Large-Scale Human-Annotated Dataset of Real-World Web Interaction Traces", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebChain aligns human web-interaction trajectories with visual, structural and action data for agent training. It also introduces WebChainBench and a mid-training approach that separates spatial grounding from planning.", "topics": ["data-generation", "evaluation", "grounding", "planning"], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2603.05295"}
{"aliases": [], "authors": ["Karan Gupta", "Pranav Vajreshwari", "Yash Pandya", "Raghav Magazine", "Akshay Nambi", "Ahmed Awadallah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "85418b7430aabc8b1955d93f697c340467a7adc21f2386f349ca4362e4f9677d", "submitted_at": "2026-03-05T20:29:12Z", "updated_at": "2026-03-05T20:29:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.06713v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.06713", "identifiers": {"arxiv": "2603.06713"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.06713"}], "name": "Scaling Agentic Capabilities, Not Context: Efficient Reinforcement Finetuning for Large Toolspaces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-05", "url": "https://arxiv.org/abs/2603.06713"}
{"aliases": [], "authors": ["Yunteng Tan", "Zhi Gao", "Xinxiao Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e9be07c38eb6bff2190126c61978d7202d7982a669f19ba7241fa6c3b6e54c71", "submitted_at": "2026-03-07T04:07:47Z", "updated_at": "2026-03-07T04:07:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.07024v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.07024", "identifiers": {"arxiv": "2603.07024"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.07024"}], "name": "Enhancing Web Agents with a Hierarchical Memory Tree", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-07", "url": "https://arxiv.org/abs/2603.07024"}
{"aliases": [], "authors": ["Lei Yin", "Wentao Cheng", "Zhida Qin", "Tianyu Huang", "Yidong Li", "Gangyi Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0c93565af5901de9c2ba10894055cbb5017ec11d31032047619553bdcc2422a0", "submitted_at": "2026-03-07T08:36:58Z", "updated_at": "2026-04-08T05:28:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.07106v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.07106", "identifiers": {"arxiv": "2603.07106"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.07106"}], "name": "AutoUE: Automated Generation of 3D Games in Unreal Engine via Multi-Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-08", "url": "https://arxiv.org/abs/2603.07106"}
{"aliases": [], "authors": ["Li Gu", "Zihuan Jiang", "Zhixiang Chi", "Huan Liu", "Ziqiang Wang", "Yuanhao Yu", "Glen Berseth", "Yang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68350312fecc36630ecaeec18a203e692e072d31ff80f442662d202b43ead788", "submitted_at": "2026-03-08T03:08:11Z", "updated_at": "2026-03-08T03:08:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.07432v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2603.07432", "identifiers": {"arxiv": "2603.07432"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.07432"}], "name": "Generalization in Online Reinforcement Learning for Mobile Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AndroidWorld-Generalization with separate tests for unseen task instances, templates, and applications. A GRPO training system combines containerized environments with asynchronous trajectory collection to study these generalization regimes.", "topics": ["reinforcement-learning", "evaluation", "robustness"], "updated_at": "2026-03-08", "url": "https://arxiv.org/abs/2603.07432"}
{"aliases": [], "authors": ["Pengfei Du"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b28eb9040c7f2a4356cefe6ceb2a387ce972296a6a59316d268efc2a2d609c83", "submitted_at": "2026-03-08T15:08:01Z", "updated_at": "2026-03-08T15:08:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.07670v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.07670", "identifiers": {"arxiv": "2603.07670"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.07670"}], "name": "Memory for Autonomous LLM Agents:Mechanisms, Evaluation, and Emerging Frontiers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-08", "url": "https://arxiv.org/abs/2603.07670"}
{"aliases": [], "authors": ["Hansi Zeng", "Zoey Li", "Yifan Gao", "Chenwei Zhang", "Xiaoman Pan", "Tao Yang", "Fengran Mo", "Jiacheng Lin", "Xian Li", "Jingbo Shang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4326fa405777b7a9a4898eac7724aec97946dedf5e198c656ae0d52a80477a61", "submitted_at": "2026-03-09T00:05:29Z", "updated_at": "2026-03-09T00:05:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.07853v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.07853", "identifiers": {"arxiv": "2603.07853"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.07853"}], "name": "SynPlanResearch-R1: Encouraging Tool Exploration for Deep Research with Synthetic Plans", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-09", "url": "https://arxiv.org/abs/2603.07853"}
{"aliases": [], "authors": ["Jiateng Liu", "Zhenhailong Wang", "Rushi Wang", "Bingxuan Li", "Jeonghwan Kim", "Aditi Tiwari", "Pengfei Yu", "Denghui Zhang", "Heng Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77eb68f31d616b8e542dbb41ada67f28bdc21f29f7da3f04a0077c0fa2d10196", "submitted_at": "2026-03-09T05:27:56Z", "updated_at": "2026-03-09T05:27:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.07978v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.07978", "identifiers": {"arxiv": "2603.07978"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.07978"}], "name": "OSExpert: Computer-Use Agents Learning Professional Skills via Exploration", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses GUI-based depth-first exploration to verify elementary software functions, assemble action primitives, and construct curricula of composed skills. OSExpert-Eval examines how the resulting procedural knowledge supports complex computer-use tasks.", "topics": ["planning", "data-generation", "evaluation"], "updated_at": "2026-03-09", "url": "https://arxiv.org/abs/2603.07978"}
{"aliases": [], "authors": ["Yuxiang Chai", "Shunye Tang", "Han Xiao", "Rui Liu", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7334e01b1fbf5cdd02c275c73d2a1b5c2931718d6019fcf6a0ea72b449704d94", "submitted_at": "2026-03-09T06:41:32Z", "updated_at": "2026-03-09T06:41:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.08013v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.08013", "identifiers": {"arxiv": "2603.08013"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.08013"}], "name": "PIRA-Bench: A Transition from Reactive GUI Agents to GUI-based Proactive Intent Recommendation Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "PIRA-Bench evaluates proactive intent recommendations from noisy GUI screenshot sequences containing interleaved tasks and user context. Its PIRF baseline combines memory and state tracking to maintain multiple task threads and identify opportunities to assist.", "topics": ["evaluation", "screen-understanding", "memory", "state-tracking", "human-in-the-loop"], "updated_at": "2026-03-09", "url": "https://arxiv.org/abs/2603.08013"}
{"aliases": [], "authors": ["Junxian Li", "Tu Lan", "Haozhen Tan", "Yan Meng", "Haojin Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c051a0cf72638e060636630f1778d452cc7625b746f6ce81265e093110c7931c", "submitted_at": "2026-03-09T12:38:28Z", "updated_at": "2026-07-01T07:09:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.08316v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.08316", "identifiers": {"arxiv": "2603.08316"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.08316"}], "name": "SlowBA: An efficiency backdoor attack towards VLM-based GUI agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-01", "url": "https://arxiv.org/abs/2603.08316"}
{"aliases": [], "authors": ["Yiping Xie", "Song Chen", "Jingxuan Xing", "Wei Jiang", "Zekun Zhu", "Yingyao Wang", "Pi Bu", "Jun Song", "Yuning Jiang", "Bo Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db31c6ca303f0c7dc43f24dd1eea69621b1e557af33122386945154e8236ab78", "submitted_at": "2026-03-09T16:04:08Z", "updated_at": "2026-03-31T03:08:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.08533v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.08533", "identifiers": {"arxiv": "2603.08533"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.08533"}], "name": "SecAgent: Efficient Mobile GUI Agent with Semantic Context", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. Mobile use is explicit, but the abstract does not identify a mobile operating system.", "published_at": "2026-03-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents a Chinese mobile GUI dataset and navigation benchmark together with SecAgent, which compresses previous screenshots and actions into semantic summaries. The agent is trained with supervised and reinforcement fine-tuning to use this compact interaction context.", "topics": ["mobile-navigation", "memory", "reinforcement-learning"], "updated_at": "2026-03-31", "url": "https://arxiv.org/abs/2603.08533"}
{"aliases": [], "authors": ["Weize Liu", "Minghui Liu", "Sy-Tuyen Ho", "Yongkyun Lee", "Andrew Adams Schoen", "Souradip Chakraborty", "Xiyao Wang", "Furong Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "335fc18afd76688fddb72433025037cbf53a8345e720f58777030a2654d1f3d2", "submitted_at": "2026-03-09T17:58:56Z", "updated_at": "2026-10-08T17:18:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.08706v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.08706", "identifiers": {"arxiv": "2603.08706"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.08706"}], "name": "Agentic Critical Training", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2603.08706"}
{"aliases": [], "authors": ["Rui Liu", "Tao Zhe", "Dongjie Wang", "Zijun Yao", "Kunpeng Liu", "Yanjie Fu", "Huan Liu", "Jian Pei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac52a18c24656e0605932488bae6d130207c888bb46773e5161cb6bab1e1d60a", "submitted_at": "2026-03-09T21:13:52Z", "updated_at": "2026-03-11T20:38:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.08938v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2603.08938", "identifiers": {"arxiv": "2603.08938"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.08938"}], "name": "AgentOS: From Application Silos to a Natural Language-Driven Data Ecosystem", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-11", "url": "https://arxiv.org/abs/2603.08938"}
{"aliases": [], "authors": ["Xiaoxing Wang", "Ning Liao", "Shikun Wei", "Chen Tang", "Feiyu Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2499301ec0d7492c80b1ebb279188fda5587dabc38b8efdd4a2bcccc1bc8f53", "submitted_at": "2026-03-10T14:23:49Z", "updated_at": "2026-03-10T14:23:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.09716v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.09716", "identifiers": {"arxiv": "2603.09716"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.09716"}], "name": "AutoAgent: Evolving Cognition and Elastic Memory Orchestration for Adaptive Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-10", "url": "https://arxiv.org/abs/2603.09716"}
{"aliases": [], "authors": ["Srivatsa Kundurthy", "Clara Na", "Michael Handley", "Zach Kirshner", "Chen Bo Calvin Zhang", "Manasi Sharma", "Emma Strubell", "John Ling"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0edd53fd833022d03e0845f1dadd99a425fd0ad4e84ee3b2c2a3ff858bd7342", "submitted_at": "2026-02-16T14:24:36Z", "updated_at": "2026-07-02T20:32:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.10002v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.10002", "identifiers": {"arxiv": "2603.10002"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.10002"}], "name": "SpreadsheetArena: Decomposing Preference in LLM Generation of Spreadsheet Workbooks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-02", "url": "https://arxiv.org/abs/2603.10002"}
{"aliases": [], "authors": ["Linxin Song", "Jieyu Zhang", "Huanxin Sheng", "Taiwei Shi", "Gupta Rahul", "Yang Liu", "Ranjay Krishna", "Jian Kang", "Jieyu Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "22b3001540ac8cc75117cdfb4f381462fe7df3fe7700a86e14b9f522803316ac", "submitted_at": "2026-03-10T19:17:22Z", "updated_at": "2026-03-10T19:17:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10178v1"}}, "environments": ["desktop", "linux", "macos", "windows", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.10178", "identifiers": {"arxiv": "2603.10178"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.10178"}], "name": "Video-Based Reward Modeling for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds ExeVR-53k from execution-video, instruction and reward triplets, including negative instructions with step annotations. A reward model uses spatiotemporal token pruning to judge task completion from the instruction and execution keyframes.", "topics": ["reward-models", "video", "evaluation"], "updated_at": "2026-03-10", "url": "https://arxiv.org/abs/2603.10178"}
{"aliases": [], "authors": ["Syed Yusuf Ahmed", "Shiwei Feng", "Chanwoo Bae", "Calix Barrus Xiangyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ad18d7fe815f5529fef663130b7ce8841763589e7f4319c0c7556ec2df1da03", "submitted_at": "2026-03-10T22:56:03Z", "updated_at": "2026-03-10T22:56:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10268v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2603.10268", "identifiers": {"arxiv": "2603.10268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.10268"}], "name": "SpecOps: A Fully Automated AI Agent Testing Framework in Real-World GUI Environments", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "SpecOps divides testing of GUI-based agents into test generation, environment setup, execution and validation handled by specialist agents. Its evaluation spans five real-world agents, including web applications and browser extensions, and measures execution reliability and bug discovery.", "topics": ["automation", "evaluation", "planning"], "updated_at": "2026-03-10", "url": "https://arxiv.org/abs/2603.10268"}
{"aliases": [], "authors": ["Sibo Zhu", "Wenyi Wu", "Kun Zhou", "Stephen Wang", "Biwei Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03b11f6c82bc8522b6450ca5a64b094ba4a7319a96f897124b3a31fa9c4f7472", "submitted_at": "2026-03-11T00:17:44Z", "updated_at": "2026-03-11T00:17:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10291v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.10291", "identifiers": {"arxiv": "2603.10291"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.10291"}], "name": "Hybrid Self-evolving Structured Memory for GUI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-11", "url": "https://arxiv.org/abs/2603.10291"}
{"aliases": [], "authors": ["Hyungjoo Chae", "Jungsoo Park", "Alan Ritter"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "178b2140d0c70ded39fede20bc91d1459d0a4bf65c5cb6a63c09c9db73828513", "submitted_at": "2026-03-11T07:58:53Z", "updated_at": "2026-03-11T07:58:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10505v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.10505", "identifiers": {"arxiv": "2603.10505"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.10505"}], "name": "Safe and Scalable Web Agent Learning via Recreated Websites", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The abstract promises a release upon acceptance; this annotation does not verify artifact availability.", "published_at": "2026-03-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces VeriEnv, which recreates websites as executable training environments and exposes internal state through a Python SDK for task generation and deterministic reward checks. Experiments examine web-agent learning across new and expanded site environments.", "topics": ["data-generation", "sandbox", "reinforcement-learning"], "updated_at": "2026-03-11", "url": "https://arxiv.org/abs/2603.10505"}
{"aliases": [], "authors": ["Marta Sumyk", "Oleksandr Kosovan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "048dc7332372585872f5fd3da47a10bd999c7520b2999e45ea6f578b7c856400", "submitted_at": "2026-03-11T09:28:41Z", "updated_at": "2026-03-12T07:30:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.10577v2"}}, "environments": ["desktop", "macos", "windows", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.10577", "identifiers": {"arxiv": "2603.10577"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.10577"}], "name": "CUAAudit: Meta-Evaluation of Vision-Language Models as Auditors of Autonomous Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "CUAAudit examines vision-language models that judge computer-use task completion from an instruction and final environment state. Across macOS, Windows and Linux benchmarks, it measures accuracy, confidence calibration and disagreement between auditors.", "topics": ["evaluation", "observability", "robustness"], "updated_at": "2026-03-12", "url": "https://arxiv.org/abs/2603.10577"}
{"aliases": [], "authors": ["Alexandre De Masi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5944045ee4562641ee02612969d1c2d038e019fa02cc8d48c671c1d5397557ad", "submitted_at": "2026-03-11T11:24:31Z", "updated_at": "2026-03-11T11:24:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10664v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.10664", "identifiers": {"arxiv": "2603.10664"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.10664"}], "name": "Terminal Is All You Need: Design Properties for Human-AI Agent Collaboration", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent conceptual HCI work about agent-facing interfaces; the abstract does not report GUI task experiments or a new evaluated CUA.", "published_at": "2026-03-11", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "This position paper derives interface design principles from terminal-based human-agent collaboration: representational compatibility, transparent actions and low participation barriers. It argues that graphical and other agent-facing modalities should provide these properties explicitly.", "topics": ["human-in-the-loop", "ui-design", "observability", "tool-integration"], "updated_at": "2026-03-11", "url": "https://arxiv.org/abs/2603.10664"}
{"aliases": [], "authors": ["Amit Singh", "Vedant Nipane", "Pulkit Agrawal", "Jatin Kishnani", "Sairanjan Mishra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "96b5cac787a40a3d059058c2e7b1b023f7721da0b1b04f241260ad0c300b180b", "submitted_at": "2026-03-11T16:32:14Z", "updated_at": "2026-03-13T15:09:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.11139v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.11139", "identifiers": {"arxiv": "2603.11139"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.11139"}], "name": "H2LooP Spark Preview: Continual Pretraining of Large Language Models for Low-Level Embedded Systems Code", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-13", "url": "https://arxiv.org/abs/2603.11139"}
{"aliases": [], "authors": ["Ziyao Wang", "Chen Chen", "Jingtao Li", "Weiming Zhuang", "Jiabo Huang", "Ang Li", "Lingjuan Lyu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cf3d647642505ae0854a48852222605018646f097b36aef97bd55e7a8cab8c75", "submitted_at": "2026-03-11T21:27:15Z", "updated_at": "2026-03-11T21:27:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.11320v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.11320", "identifiers": {"arxiv": "2603.11320"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.11320"}], "name": "UniCompress: Token Compression for Unified Vision-Language Understanding and Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-11", "url": "https://arxiv.org/abs/2603.11320"}
{"aliases": [], "authors": ["Ching-Yu Kao", "Xinfeng Li", "Shenyu Dai", "Tianze Qiu", "Pengcheng Zhou", "Eric Hanchen Jiang", "Philip Sperl"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "919a860952ce5fbb15c1bef9a5ef3766e46a15d3f4a504d10caa111ba0d9826a", "submitted_at": "2026-03-12T12:35:46Z", "updated_at": "2026-03-12T12:35:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.11862v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2603.11862", "identifiers": {"arxiv": "2603.11862"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.11862"}], "name": "You Told Me to Do It: Measuring Instructional Text-induced Private Data Leakage in LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-12", "url": "https://arxiv.org/abs/2603.11862"}
{"aliases": [], "authors": ["Rui Shao", "Ruize Gao", "Bin Xie", "Yixing Li", "Kaiwen Zhou", "Shuai Wang", "Weili Guan", "Gongwei Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6421a7928b502cd431964e6f84a1a58d4cd16dbbf1d9b3d1ecc1c2b3dae863bf", "submitted_at": "2026-03-12T16:40:59Z", "updated_at": "2026-03-12T16:40:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.12138v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.12138", "identifiers": {"arxiv": "2603.12138"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.12138"}], "name": "HATS: Hardness-Aware Trajectory Synthesis for GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces HATS, which directs GUI trajectory exploration toward semantically ambiguous actions and iteratively checks instruction-execution alignment. Feedback from trajectory refinement updates the exploration signal used to collect subsequent training examples.", "topics": ["data-generation", "robustness", "action-prediction"], "updated_at": "2026-03-12", "url": "https://arxiv.org/abs/2603.12138"}
{"aliases": [], "authors": ["Aadyaa Maddi", "Prakhar Naval", "Deepti Mande", "Shane Duan", "Muckai Girish", "Vyas Sekar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8feb29743ea06d6d3d5750b50b6933345f67c30ab9cfea0fd0756ec643150f86", "submitted_at": "2026-03-12T22:01:36Z", "updated_at": "2026-03-12T22:01:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.12483v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.12483", "identifiers": {"arxiv": "2603.12483"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.12483"}], "name": "Generating Expressive and Customizable Evals for Timeseries Data Analysis Agents with AgentFuel", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-12", "url": "https://arxiv.org/abs/2603.12483"}
{"aliases": [], "authors": ["Orit Shahnovsky", "Rotem Dror"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2442ff48e3a112401c9f8910d904d800432a9760cc35aa3615670f9bf52b7ef2", "submitted_at": "2026-03-13T06:46:32Z", "updated_at": "2026-03-13T06:46:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.12710v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.12710", "identifiers": {"arxiv": "2603.12710"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.12710"}], "name": "AI Planning Framework for LLM-Based Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-13", "url": "https://arxiv.org/abs/2603.12710"}
{"aliases": [], "authors": ["Xunzhuo Liu", "Bowei He", "Xue Liu", "Andy Luo", "Haichen Zhang", "Huamin Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "274deff6cacfb5f1b7b620a86abe7ef6496b673d929de39376a9ad42dd3c4c02", "submitted_at": "2026-03-13T09:21:25Z", "updated_at": "2026-03-13T09:21:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.12823v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.12823", "identifiers": {"arxiv": "2603.12823"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.12823"}], "name": "Adaptive Vision-Language Model Routing for Computer Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-03-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adaptive VLM Routing estimates GUI-action difficulty and small-model confidence to select a model under a predicted reliability threshold. Grounding and agent-routing benchmarks assess the cost–accuracy trade-off, with prior interaction context used for warm agents.", "topics": ["efficiency", "grounding", "memory", "evaluation"], "updated_at": "2026-03-13", "url": "https://arxiv.org/abs/2603.12823"}
{"aliases": [], "authors": ["Chenlong Yin", "Runpeng Geng", "Yanting Wang", "Jinyuan Jia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7a4e69b71c786405ef3b4b4580eabdecac799035ef5fcc423215f143f500c500", "submitted_at": "2026-03-13T14:34:54Z", "updated_at": "2026-07-23T13:46:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.13026v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.13026", "identifiers": {"arxiv": "2603.13026"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.13026"}], "name": "PISmith: Reinforcement Learning-based Red Teaming for Prompt Injection Defenses", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-23", "url": "https://arxiv.org/abs/2603.13026"}
{"aliases": [], "authors": ["Florin Adrian Chitan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eb1725f42fe7434d70177dcf1da1761aa0c0f3675e05cac039ea021d629066e4", "submitted_at": "2026-02-22T12:25:42Z", "updated_at": "2026-02-22T12:25:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.13247v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.13247", "identifiers": {"arxiv": "2603.13247"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.13247"}], "name": "ILION: Deterministic Pre-Execution Safety Gates for Agentic AI Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-22", "url": "https://arxiv.org/abs/2603.13247"}
{"aliases": [], "authors": ["Mohamed Aghzal", "Gregory J. Stein", "Ziyu Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "71bd0cbadafc5171fe0f46ba74f9764d29fba590f0d24d49a6f9eb6ca5be1cf4", "submitted_at": "2026-03-15T06:52:06Z", "updated_at": "2026-04-27T17:57:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.14248v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.14248", "identifiers": {"arxiv": "2603.14248"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.14248"}], "name": "Why Do LLM-based Web Agents Fail? A Hierarchical Planning Perspective", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-27", "url": "https://arxiv.org/abs/2603.14248"}
{"aliases": [], "authors": ["Ziwei Liu", "Tao Feng", "Borui Kang", "Yanbing Yang", "Jun Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4507a652e2b3c3857289763442fffd8db21726b97fb7b6304ad61a2139ef5af", "submitted_at": "2026-03-15T15:47:47Z", "updated_at": "2026-03-15T15:47:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.14448v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.14448", "identifiers": {"arxiv": "2603.14448"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.14448"}], "name": "Zoom to Essence: Trainless GUI Grounding by Inferring upon Interface Elements", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents ZoomUI, a training-free grounding method that expresses instructions as visual element descriptions and uses model attention to zoom progressively into relevant interface regions. Evaluation examines instruction-to-element localization on GUI benchmarks.", "topics": ["grounding", "training-free", "screen-understanding"], "updated_at": "2026-03-15", "url": "https://arxiv.org/abs/2603.14448"}
{"aliases": [], "authors": ["Xunzhuo Liu", "Bowei He", "Xue Liu", "Andy Luo", "Haichen Zhang", "Huamin Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c7cb0e4784c2a58feeca27a1687d1f73880d9713c16387f0bbff9ed34c427c7", "submitted_at": "2026-03-16T01:31:32Z", "updated_at": "2026-03-16T01:31:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.14707v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.14707", "identifiers": {"arxiv": "2603.14707"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.14707"}], "name": "Visual Confused Deputy: Exploiting and Defending Perception Failures in Computer-Using Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Formalizes a visual confused-deputy failure in which a computer agent acts on a misperceived UI object. A separate guardrail checks both the visual target and the agent's action reasoning against deployment-specific knowledge before permitting execution.", "topics": ["safety-security", "grounding", "robustness"], "updated_at": "2026-03-16", "url": "https://arxiv.org/abs/2603.14707"}
{"aliases": [], "authors": ["Yang Li", "Yuchen Liu", "Haoyu Lu", "Zhiqiang Xia", "Hongzhen Wang", "Kaiyang Han", "Changpeng Yang", "Jinyang Wu", "Jiaming Xu", "Runyu Shi", "Ying Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf6c3df80e405968d9ae2353639f3e039d3fcf8351d89db212bda038e3dfac9b", "submitted_at": "2026-03-16T09:45:33Z", "updated_at": "2026-03-16T09:45:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.15039v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.15039", "identifiers": {"arxiv": "2603.15039"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.15039"}], "name": "GUI-CEval: A Hierarchical and Comprehensive Chinese Benchmark for Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-CEval evaluates Chinese mobile GUI agents through a hierarchy of atomic abilities and application-level tasks collected on physical devices. Its diagnostics cover perception, planning, reflection, execution and post-action evaluation across a broad set of apps.", "topics": ["evaluation", "mobile-navigation", "self-reflection", "screen-understanding"], "updated_at": "2026-03-16", "url": "https://arxiv.org/abs/2603.15039"}
{"aliases": [], "authors": ["Mateusz Dziemian", "Maxwell Lin", "Xiaohan Fu", "Micha Nowak", "Nick Winter", "Eliot Jones", "Andy Zou", "Lama Ahmad", "Kamalika Chaudhuri", "Sahana Chennabasappa", "Xander Davies", "Lauren Deason", "Benjamin L. Edelman", "Tanner Emek", "Ivan Evtimov", "Jim Gust", "Maia Hamin", "Kat He", "Klaudia Krawiecka", "Riccardo Patana", "Neil Perry", "Troy Peterson", "Xiangyu Qi", "Javier Rando", "Zifan Wang", "Zihan Wang", "Spencer Whitman", "Eric Winsor", "Arman Zharmagambetov", "Matt Fredrikson", "Zico Kolter"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7fe52247e01f331021013dbc3003b3b828e134d7706ec1f587210ad286eee1d", "submitted_at": "2026-03-16T14:49:36Z", "updated_at": "2026-03-16T14:49:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.15714v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.15714", "identifiers": {"arxiv": "2603.15714"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.15714"}], "name": "How Vulnerable Are AI Agents to Indirect Prompt Injections? Insights from a Large-Scale Public Competition", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-03-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "This public red-teaming study evaluates indirect prompt injection that both changes agent behavior and hides the compromise from the final response. Its scenarios include computer use alongside tool calling and coding, with analyses of transfer and model-dependent vulnerability.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-03-16", "url": "https://arxiv.org/abs/2603.15714"}
{"aliases": [], "authors": ["Zhengbo Zhang", "Jinbo Su", "Zhaowen Zhou", "Changtao Miao", "Yuhan Hong", "Qimeng Wu", "Yumeng Liu", "Feier Wu", "Yihe Tian", "Yuhao Liang", "Zitong Shan", "Wanke Xia", "Yi-Fan Zhang", "Bo Zhang", "Zhe Li", "Shiming Xiang", "Ying Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c7a59738348defc6f8cb4583b7a1d1390a764dfa8d7da914ed33d43e6041c48", "submitted_at": "2026-03-17T09:24:13Z", "updated_at": "2026-03-18T15:40:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.16289v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.16289", "identifiers": {"arxiv": "2603.16289"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.16289"}], "name": "VisBrowse-Bench: Benchmarking Visual-Native Search for Multimodal Browsing Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-18", "url": "https://arxiv.org/abs/2603.16289"}
{"aliases": [], "authors": ["Karthik Ragunath Ananda Kumar", "Subrahmanyam Arunachalam"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1cbb8cd78f0e580e79433e197e590cda661f4b13a2d4c20ac14ebe612026f67b", "submitted_at": "2026-03-17T17:45:53Z", "updated_at": "2026-03-17T17:45:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.16839v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.16839", "identifiers": {"arxiv": "2603.16839"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.16839"}], "name": "Learning to Present: Inverse Specification Rewards for Agentic Slide Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-17", "url": "https://arxiv.org/abs/2603.16839"}
{"aliases": [], "authors": ["Ziyang Cai", "Amir Saeidi", "Harkirat Behl"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f9bd5816022050509f226390933d35247f4d70ee69f2cf4d4190f06465633301", "submitted_at": "2026-03-17T23:43:16Z", "updated_at": "2026-08-16T18:47:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.17216v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.17216", "identifiers": {"arxiv": "2603.17216"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.17216"}], "name": "ML-AutoResearch: Training Machine Learning Research Agents with Automatically Generated Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-16", "url": "https://arxiv.org/abs/2603.17216"}
{"aliases": [], "authors": ["Nathan Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3b429d277ec776af18977a72aaeb95000a4fedf66aa89c0366ab02120d4cdbe", "submitted_at": "2026-03-18T04:41:16Z", "updated_at": "2026-03-18T04:41:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.17357v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.17357", "identifiers": {"arxiv": "2603.17357"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.17357"}], "name": "WebPII: Benchmarking Visual PII Detection for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebPII provides synthetic e-commerce screenshots with fine-grained annotations for sensitive information, including partially filled forms and transaction identifiers. It also trains WebRedact to detect visual PII as a privacy component for computer-use agents.", "topics": ["evaluation", "data-generation", "safety-security", "screen-understanding"], "updated_at": "2026-03-18", "url": "https://arxiv.org/abs/2603.17357"}
{"aliases": [], "authors": ["Seyed Mohammad Asghari", "Chris Chute", "Vikranth Dwaracherla", "Xiuyuan Lu", "Mehdi Jafarnia", "Victor Minden", "Zheng Wen", "Benjamin Van Roy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "692cbfbee3906a32744cd1e90c7aa3cb74359815f61b8ad4ac2fd56ea30c770f", "submitted_at": "2026-03-18T05:47:59Z", "updated_at": "2026-08-20T20:29:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.17378v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.17378", "identifiers": {"arxiv": "2603.17378"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.17378"}], "name": "Efficient Exploration at Scale", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-20", "url": "https://arxiv.org/abs/2603.17378"}
{"aliases": [], "authors": ["Siqi Pei", "Liang Tang", "Tiaonan Duan", "Long Chen", "Shuxian Li", "Kaer Huang", "Yanzhe Jing", "Yiqiang Yan", "Bo Zhang", "Chenghao Jiang", "Borui Zhang", "Jiwen Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2101a32746d6c668bff79461983b2e12138959f49ab3cfcb542ecbd629317f1d", "submitted_at": "2026-03-18T07:26:18Z", "updated_at": "2026-03-18T07:26:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.17441v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.17441", "identifiers": {"arxiv": "2603.17441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.17441"}], "name": "AdaZoom-GUI: Adaptive Zoom-based GUI Grounding with Instruction Refinement", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines instruction rewriting with conditional second-stage zooming for small GUI targets. A grounding dataset and GRPO training support joint prediction of click coordinates and element bounding boxes.", "topics": ["grounding", "reinforcement-learning", "screen-understanding"], "updated_at": "2026-03-18", "url": "https://arxiv.org/abs/2603.17441"}
{"aliases": [], "authors": ["Ruize Ma", "Yilei Jiang", "Shilin Zhang", "Zheng Ma", "Yi Feng", "Vincent Ng", "Zhi Wang", "Xiangyu Yue", "Chuanyi Li", "Lewei Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d2427bb3c8b15d7a678389a141d3fc1b26c08e41410842e7fbc9040e3b92a25", "submitted_at": "2026-03-18T15:24:22Z", "updated_at": "2026-03-18T15:24:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.17826v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2603.17826", "identifiers": {"arxiv": "2603.17826"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.17826"}], "name": "FailureMem: A Failure-Aware Multimodal Framework for Autonomous Software Repair", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-18", "url": "https://arxiv.org/abs/2603.17826"}
{"aliases": [], "authors": ["Yibo Shi", "Jungang Li", "Linghao Zhang", "Zihao Dongfang", "Biao Wu", "Sicheng Tao", "Yibo Yan", "Chenxi Qin", "Weiting Liu", "Zhixin Lin", "Hanqian Li", "Yu Huang", "Song Dai", "Yonghua Hei", "Yue Ding", "Xiang Li", "Shikang Wang", "Chengdong Xu", "Jingqi Liu", "Xueying Ma", "Zhiwen Zheng", "Xiaofei Zhang", "Bincheng Wang", "Nichen Yang", "Jie Wu", "Lihua Tian", "Chen Li", "Xuming Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ce6da91e59ac57fa24ff991ca023b84448c54d477cacb2d6fded9bf527dd29d", "submitted_at": "2026-03-19T02:45:21Z", "updated_at": "2026-03-19T02:45:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.18429v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.18429", "identifiers": {"arxiv": "2603.18429"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.18429"}], "name": "AndroTMem: From Interaction Trajectories to Anchored Memory in Long-Horizon GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AndroTMem-Bench for Android tasks whose later actions depend on earlier UI states. Anchored State Memory retains causally linked intermediate states for subgoal-specific retrieval and evaluates them against full-history and summary-based memory.", "topics": ["memory", "long-horizon", "evaluation"], "updated_at": "2026-03-19", "url": "https://arxiv.org/abs/2603.18429"}
{"aliases": [], "authors": ["Haochen Zhao", "Shaoyang Cui"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0beb3fe342f080f0727181aba55fc1167d2512dd04542986448c5006b6b8213b", "submitted_at": "2026-03-19T11:14:45Z", "updated_at": "2026-03-19T11:14:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.18762v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2603.18762", "identifiers": {"arxiv": "2603.18762"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.18762"}], "name": "ClawTrap: A MITM-Based Red-Teaming Framework for Real-World OpenClaw Security Evaluation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents a man-in-the-middle network red-teaming framework for evaluating autonomous OpenClaw agents under live traffic manipulation.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2026-03-19", "url": "https://arxiv.org/abs/2603.18762"}
{"aliases": [], "authors": ["Zehao Li", "Zhenyu Wu", "Yibo Zhao", "Bowen Yang", "Jingjing Xie", "Zhaoyang Liu", "Zhoumianze Liu", "Kaiming Jin", "Jianze Liang", "Zonglin Li", "Feng Wu", "Bowen Zhou", "Zun Wang", "Zichen Ding"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff31371580af6a0c43a6c26f076d5060ee06f974d3a6c1ade67348dc0af4859d", "submitted_at": "2026-03-19T17:47:47Z", "updated_at": "2026-03-19T17:47:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19191v1"}}, "environments": ["android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2603.19191", "identifiers": {"arxiv": "2603.19191"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.19191"}], "name": "OS-Themis: A Scalable Critic Framework for Generalist GUI Rewards", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents OS-Themis, a multi-agent critic that divides GUI trajectories into verifiable milestones and reviews their evidence before assigning an outcome. OGRBench evaluates GUI rewards across platforms, with AndroidWorld experiments examining training and trajectory filtering.", "topics": ["reward-models", "reinforcement-learning", "evaluation"], "updated_at": "2026-03-19", "url": "https://arxiv.org/abs/2603.19191"}
{"aliases": [], "authors": ["Kai Wang", "Haoyang You", "Yang Zhang", "Zhongjie Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5d1781b852b97be824a5af1d943f0cc522dbcabcf1af17626628c35309f79edb", "submitted_at": "2026-03-14T07:12:42Z", "updated_at": "2026-03-14T07:12:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19313v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.19313", "identifiers": {"arxiv": "2603.19313"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.19313"}], "name": "Memory-Driven Role-Playing: Evaluation and Enhancement of Persona Knowledge Utilization in LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-14", "url": "https://arxiv.org/abs/2603.19313"}
{"aliases": [], "authors": ["Toan Tran", "Olivera Kotevska", "Li Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b64315ec368626e3a4d1de41fac9dfbc7571a8af9447136a71ecf508171835d1", "submitted_at": "2026-03-19T18:10:18Z", "updated_at": "2026-09-16T23:23:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.19375v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.19375", "identifiers": {"arxiv": "2603.19375"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.19375"}], "name": "Automated Membership Inference Attacks (AutoMIA): Discovering MIA Signal Computations using LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-16", "url": "https://arxiv.org/abs/2603.19375"}
{"aliases": [], "authors": ["Taiyi Wang", "Sian Gooding", "Florian Hartmann", "Oriana Riva", "Edward Grefenstette"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5538933a0e3d11df6ff4d767849979703d83412599fbbb0f8b432c4d9850633b", "submitted_at": "2026-03-20T06:32:16Z", "updated_at": "2026-03-20T06:32:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19685v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.19685", "identifiers": {"arxiv": "2603.19685"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.19685"}], "name": "A Subgoal-driven Framework for Improving Long-Horizon LLM Agents", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-03-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "The framework separates inference-time subgoal planning from MiRA, a reinforcement-learning method that supplies milestone rewards. WebArena-Lite experiments evaluate these mechanisms for long-horizon web navigation.", "topics": ["planning", "reinforcement-learning", "reward-models", "long-horizon", "evaluation"], "updated_at": "2026-03-20", "url": "https://arxiv.org/abs/2603.19685"}
{"aliases": [], "authors": ["Rozain Shakeel", "Abdul Rahman Mohammad Ali", "Muneeb Mushtaq", "Tausifa Jan Saleem", "Tajamul Ashraf"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cfa6cd9aead44150fdc997654ded7253b2890e936b94e44be2fb33642b24d302", "submitted_at": "2026-03-20T14:43:53Z", "updated_at": "2026-03-20T14:43:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19993v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2603.19993", "identifiers": {"arxiv": "2603.19993"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.19993"}], "name": "MedSPOT: A Workflow-Aware Sequential Grounding Benchmark for Clinical GUI", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2026-03-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Benchmarks sequential target localization in clinical software workflows, emphasizing how early mistakes affect subsequent interaction steps.", "topics": ["grounding", "evaluation"], "updated_at": "2026-03-20", "url": "https://arxiv.org/abs/2603.19993"}
{"aliases": [], "authors": ["Zhuofeng Li", "Dongfu Jiang", "Xueguang Ma", "Haoxiang Zhang", "Ping Nie", "Yuyu Zhang", "Kai Zou", "Jianwen Xie", "Yu Zhang", "Wenhu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bec4e8a0a3bc989ff2eab639473d8d03c3bb4e2243c14a9c2aa278a69a56896", "submitted_at": "2026-03-17T20:10:12Z", "updated_at": "2026-09-10T09:50:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.20278v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2603.20278", "identifiers": {"arxiv": "2603.20278"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.20278"}], "name": "OpenResearcher: A Fully Open Pipeline for Long-Horizon Deep Research Trajectory Synthesis", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-10", "url": "https://arxiv.org/abs/2603.20278"}
{"aliases": [], "authors": ["Zijian Lu", "Yiping Zuo", "Yupeng Nie", "Xin He", "Weibei Fan", "Lianyong Qi", "Shi Jin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e8182e64c7a57a2dd6eb381762bc9a60b08e13ba5a337d0e9125ab6a5518b6c", "submitted_at": "2026-03-20T09:25:50Z", "updated_at": "2026-04-15T07:34:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.20340v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.20340", "identifiers": {"arxiv": "2603.20340"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.20340"}], "name": "ContractSkill: Repairable Contract-Based Skills for Multimodal Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The abstract limits its portability evidence to transfer within a benchmark family, not generalization across arbitrary benchmarks.", "published_at": "2026-03-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Turns draft web-agent skills into structured executable artifacts that support deterministic checks, fault localization and local repair. VisualWebArena and MiniWoB experiments examine execution and reuse within the same benchmark family.", "topics": ["web-navigation", "planning", "robustness"], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2603.20340"}
{"aliases": [], "authors": ["Yizhe Zhao", "Yongjian Fu", "Zihao Feng", "Hao Pan", "Yongheng Deng", "Yaoxue Zhang", "Ju Ren"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f25f4999fc9d20433f96e22651311cccf78b5e2ad7cbc5578fe1765ac8be6a00", "submitted_at": "2026-03-20T12:19:41Z", "updated_at": "2026-03-20T12:19:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.20351v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2603.20351", "identifiers": {"arxiv": "2603.20351"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.20351"}], "name": "MANA: Towards Efficient Mobile Ad Detection via Multimodal Agentic UI Navigation", "notes": "Reviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.", "published_at": "2026-03-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "MANA combines visual, temporal and exploration signals to guide navigation toward mobile advertisements. It evaluates ad detection and interaction cost on commercial smartphones, including obfuscated-ad cases.", "topics": ["mobile-navigation", "screen-understanding", "efficiency", "safety-security"], "updated_at": "2026-03-20", "url": "https://arxiv.org/abs/2603.20351"}
{"aliases": [], "authors": ["Ruiying Ma", "Shreya Shankar", "Ruiqi Chen", "Yiming Lin", "Sepanta Zeighami", "Rajoshi Ghosh", "Abhinav Gupta", "Anushrut Gupta", "Tanmai Gopal", "Aditya G. Parameswaran"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b6888d9fd47370a817f2ad74affc172a6ddc80607b8a192d1ef672f25eaafb1", "submitted_at": "2026-03-21T00:14:53Z", "updated_at": "2026-10-07T20:17:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.20576v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2603.20576", "identifiers": {"arxiv": "2603.20576"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.20576"}], "name": "Can AI Agents Answer Your Data Questions? A Benchmark for Data Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-07", "url": "https://arxiv.org/abs/2603.20576"}
{"aliases": [], "authors": ["Jiayi Geng", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b0aefc9563b059bf26a19639955fac763cfd316e6d4decf30950fc68533407c", "submitted_at": "2026-03-23T02:26:35Z", "updated_at": "2026-07-08T16:30:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.21489v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.21489", "identifiers": {"arxiv": "2603.21489"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.21489"}], "name": "Effective Strategies for Asynchronous Software Engineering Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-08", "url": "https://arxiv.org/abs/2603.21489"}
{"aliases": [], "authors": ["Honglin He", "Yukai Ma", "Brad Squicciarini", "Wayne Wu", "Bolei Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e65c8ea83075d14f483813a0726e354dc30148de7f7d5338c1f7c8d6722f7ea", "submitted_at": "2026-03-23T19:41:21Z", "updated_at": "2026-03-23T19:41:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.22527v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.22527", "identifiers": {"arxiv": "2603.22527"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.22527"}], "name": "Learning Sidewalk Autopilot from Multi-Scale Imitation with Corrective Behavior Expansion", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-23", "url": "https://arxiv.org/abs/2603.22527"}
{"aliases": [], "authors": ["Shoubin Yu", "Lei Shu", "Antoine Yang", "Yao Fu", "Srinivas Sunkara", "Maria Wang", "Jindong Chen", "Mohit Bansal", "Boqing Gong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e4ed7391969359c2594e8b96b8a3a559ef98d2c84a10c1a2c39b3bd962b58c5", "submitted_at": "2026-03-23T19:47:58Z", "updated_at": "2026-03-23T19:47:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.22529v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.22529", "identifiers": {"arxiv": "2603.22529"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.22529"}], "name": "Ego2Web: A Web Agent Benchmark Grounded in Egocentric Videos", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Ego2Web pairs first-person videos with online tasks that require agents to connect the user’s physical context to web actions. It also evaluates an automatic task judge against human judgments and analyzes failures in video understanding and web execution.", "topics": ["evaluation", "video", "web-navigation", "planning"], "updated_at": "2026-03-23", "url": "https://arxiv.org/abs/2603.22529"}
{"aliases": [], "authors": ["Wanying Mo", "Jijia Lai", "Xiaoming Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a4616e7df202d8e0ae3185ac5625fbc68ef6e72f54d2693d6be34dd6b0693c0", "submitted_at": "2026-03-24T08:05:05Z", "updated_at": "2026-03-24T08:05:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.22917v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.22917", "identifiers": {"arxiv": "2603.22917"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.22917"}], "name": "IntentWeave: A Progressive Entry Ladder for Multi-Surface Browser Agents in Cloud Portals", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-24", "url": "https://arxiv.org/abs/2603.22917"}
{"aliases": [], "authors": ["Yutao Luo", "Haotian Zhu", "Shuchao Pang", "Zhigang Lu", "Tian Dong", "Yongbin Zhou", "Minhui Xue"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37960f42162d2bff2fca5e71e350f1ef7c4d86e3f438fd777a88dc2280d3aabe", "submitted_at": "2026-03-24T09:51:43Z", "updated_at": "2026-03-24T09:51:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.23007v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.23007", "identifiers": {"arxiv": "2603.23007"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.23007"}], "name": "AgentRAE: Remote Action Execution through Notification-based Visual Backdoors against Screenshots-based Mobile GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-03-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "AgentRAE studies notification-based visual backdoors that redirect mobile GUI-agent actions. Its training separates sensitivity to icon differences from trigger/action association, with evaluations of clean behavior and targeted redirection.", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2026-03-24", "url": "https://arxiv.org/abs/2603.23007"}
{"aliases": [], "authors": ["Yuxi Chen", "Haoyu Zhai", "Chenkai Wang", "Rui Yang", "Lingming Zhang", "Gang Wang", "Huan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2e699a4bd19fb8a0cd68ad47ee432a937323e21493da27e881b81f62c14c48ca", "submitted_at": "2026-03-23T23:35:59Z", "updated_at": "2026-06-29T04:46:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.23559v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.23559", "identifiers": {"arxiv": "2603.23559"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.23559"}], "name": "CAPTCHA Solving for Native GUI Agents: Automated Reasoning-Action Data Generation and Self-Corrective Training", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-29", "url": "https://arxiv.org/abs/2603.23559"}
{"aliases": [], "authors": ["Yi Han", "Yan Wang", "Lingfei Qian", "Haohang Li", "Yupeng Cao", "Yueru He", "Xueqing Peng", "Nanhan Shen", "Yitao Xu", "Yankai Chen", "Dongji Feng", "Jimin Huang", "Xue Liu", "Jian-Yun Nie", "Sophia Ananiadou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "04b593ac719771da0d2bba126cfc48bb9dc6cab62dfcbb6b587a7f501e72adec", "submitted_at": "2026-03-24T18:25:00Z", "updated_at": "2026-05-16T06:55:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.23638v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.23638", "identifiers": {"arxiv": "2603.23638"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.23638"}], "name": "Can LLM Agents Be CFOs? Benchmarking Long-Horizon Resource Allocation in an Uncertain Enterprise Environment", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-16", "url": "https://arxiv.org/abs/2603.23638"}
{"aliases": [], "authors": ["Qianlong Lan", "Anuj Kaul"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb8164fa076c55cddf73508e6b68ce24a6594917e65862504ba9f08b4e97ee48", "submitted_at": "2026-03-24T23:49:15Z", "updated_at": "2026-03-24T23:49:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.23791v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2603.23791", "identifiers": {"arxiv": "2603.23791"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.23791"}], "name": "The Cognitive Firewall:Securing Browser Based AI Agents Against Indirect Prompt Injection Via Hybrid Edge Cloud Defense", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes a split-compute browser defense pairing a local visual sentinel, a cloud planner, and deterministic policy enforcement against indirect prompt injection.", "topics": ["safety-security", "web-navigation"], "updated_at": "2026-03-24", "url": "https://arxiv.org/abs/2603.23791"}
{"aliases": [], "authors": ["Akash Ghosh", "Tajamul Ashraf", "Rishu Kumar Singh", "Numan Saeed", "Sriparna Saha", "Xiuying Chen", "Salman Khan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8fb6a294616e1be46c5da05f60cda7d43c05d8a1ce71116c5e30cb94507cd56f", "submitted_at": "2026-03-25T10:25:48Z", "updated_at": "2026-03-25T10:25:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.24157v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.24157", "identifiers": {"arxiv": "2603.24157"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.24157"}], "name": "CarePilot: A Multi-Agent Framework for Long-Horizon Computer Task Automation in Healthcare", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-03-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "CareFlow benchmarks long healthcare-software workflows, while CarePilot coordinates action proposals, execution checks and corrective feedback using short- and long-term memory. Its actor consumes visual interfaces and system state to select semantic actions.", "topics": ["automation", "planning", "memory", "evaluation", "long-horizon", "workflow"], "updated_at": "2026-03-25", "url": "https://arxiv.org/abs/2603.24157"}
{"aliases": [], "authors": ["Shengcheng Yu", "Yuchen Ling", "Chunrong Fang", "Zhenyu Chen", "Chunyang Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "92c8f12416487aa3490c7a1f5ba4f5dfc3f9267ce7bee59a5494a8678d0fb60c", "submitted_at": "2026-03-25T10:29:04Z", "updated_at": "2026-04-15T12:04:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.24160v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2603.24160", "identifiers": {"arxiv": "2603.24160"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.24160"}], "name": "Towards Automated Crowdsourced Testing via Personified-LLM", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2603.24160"}
{"aliases": [], "authors": ["Xiangru Jian", "Shravan Nayak", "Kevin Qinghong Lin", "Aarash Feizi", "Kaixin Li", "Patrice Bechard", "Spandana Gella", "Sai Rajeswar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15237fda2f3799314009635017487749dbedc9b1a0ffdd6b56a8d28eaf01481d", "submitted_at": "2026-03-25T15:52:56Z", "updated_at": "2026-03-25T15:52:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.24440v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.24440", "identifiers": {"arxiv": "2603.24440"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.24440"}], "name": "CUA-Suite: Massive Human-annotated Video Demonstrations for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "CUA-Suite combines continuous expert desktop demonstrations in VideoCUA with the UI-Vision evaluation benchmark and GroundCUA grounding data. Screen recordings, cursor traces and layered annotations support study of temporal behavior and professional software workflows.", "topics": ["data-generation", "video", "evaluation", "grounding", "workflow"], "updated_at": "2026-03-25", "url": "https://arxiv.org/abs/2603.24440"}
{"aliases": [], "authors": ["Zichuan Lin", "Feiyu Liu", "Yijun Yang", "Jiafei Lyu", "Yiming Gao", "Yicheng Liu", "Zhicong Lu", "Yangbin Yu", "Mingyu Yang", "Junyou Li", "Deheng Ye", "Jie Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa03c8716edf782ce1a86e404107aa2569be63484dfedb472d3710dc82f405d0", "submitted_at": "2026-03-25T17:10:29Z", "updated_at": "2026-03-25T17:10:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.24533v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.24533", "identifiers": {"arxiv": "2603.24533"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.24533"}], "name": "UI-Voyager: A Self-Evolving GUI Agent Learning via Failed Experience", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a two-stage mobile GUI learner that first co-evolves trajectories and policies through rejection fine-tuning. Group Relative Self-Distillation then uses successful rollout branches to provide step-level supervision for failed trajectories, evaluated on AndroidWorld.", "topics": ["imitation-learning", "data-generation", "mobile-navigation"], "updated_at": "2026-03-25", "url": "https://arxiv.org/abs/2603.24533"}
{"aliases": [], "authors": ["Marc-Antoine Allard", "Arnaud Teinturier", "Victor Xing", "Gautier Viaud"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ab0dd9f2cba16f82b20cf679d67e9b3b302b2e25ed812ad22867e0b74dc0b172", "submitted_at": "2026-03-25T11:43:22Z", "updated_at": "2026-03-31T19:23:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.24639v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.24639", "identifiers": {"arxiv": "2603.24639"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.24639"}], "name": "Experiential Reflective Learning for Self-Improving LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-31", "url": "https://arxiv.org/abs/2603.24639"}
{"aliases": [], "authors": ["Fanheng Kong", "Jingyuan Zhang", "Yang Yue", "Chenxi Sun", "Yang Tian", "Shi Feng", "Xiaocui Yang", "Daling Wang", "Yu Tian", "Jun Du", "Wenchong Zeng", "Han Li", "Kun Gai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8414eb53c66f745fb074d9f2699f76ffea81cb55f98aacf9ce896401937f80c5", "submitted_at": "2026-03-26T09:27:29Z", "updated_at": "2026-03-26T09:27:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.25226v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.25226", "identifiers": {"arxiv": "2603.25226"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.25226"}], "name": "WebTestBench: Evaluating Computer-Use Agents towards End-to-End Automated Web Testing", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebTestBench for end-to-end website testing, separating checklist generation from defect detection. Its WebTester baseline evaluates test coverage, detection and sustained interaction across web application categories.", "topics": ["evaluation", "web-navigation", "automation"], "updated_at": "2026-03-26", "url": "https://arxiv.org/abs/2603.25226"}
{"aliases": [], "authors": ["Saelyne Yang", "Jaesang Yu", "Yi-Hao Peng", "Kevin Qinghong Lin", "Jae Won Cho", "Yale Song", "Juho Kim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b3dd3efca196fa3dd168431102dbb5f0f763e7d9e54b3d709ad2362a9c6f154", "submitted_at": "2026-03-26T19:37:53Z", "updated_at": "2026-03-26T19:37:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.25864v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2603.25864", "identifiers": {"arxiv": "2603.25864"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.25864"}], "name": "GUIDE: A Benchmark for Understanding and Assisting Users in Open-Ended GUI Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUIDE evaluates whether models can recognize a user’s behavior, infer intent and decide when to offer help during open-ended software tasks. It uses narrated screen recordings of novice users and studies the effect of additional user context on these assistance decisions.", "topics": ["evaluation", "human-in-the-loop", "video", "screen-understanding"], "updated_at": "2026-03-26", "url": "https://arxiv.org/abs/2603.25864"}
{"aliases": [], "authors": ["Daiqiang Li", "Zihao Pan", "Zeyu Zhang", "Xuyang Liu", "Shijia Xu", "Ronghao Chen", "Huacan Wang", "Honggang Chen", "Linfeng Zhang", "Zhangquan Chen", "Haiyun Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "738d5048bb7a137d848b6f6068a7ce92694469f1e7bce0702dce422c230a3d0b", "submitted_at": "2026-03-27T03:21:19Z", "updated_at": "2026-08-09T16:37:37Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2603.26041v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.26041", "identifiers": {"arxiv": "2603.26041"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.26041"}], "name": "Where and How to Prune: An Empirical Study of Visual Token Pruning for GUI Agent Navigation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies visual token pruning for GUI navigation, separating reusable screenshot encoder embeddings from later language-model computation. The analysis examines foreground-background token allocation and spatial coverage when pruning historical screenshots.", "topics": ["efficiency", "interaction-history", "screen-understanding"], "updated_at": "2026-08-09", "url": "https://arxiv.org/abs/2603.26041"}
{"aliases": [], "authors": ["Shrinidhi Kumbhar", "Haofu Liao", "Srikar Appalaraju", "Kunwar Yashraj Singh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0838afd6777af3e3995fa3bf9645102f98962cccd81c1db438a007b4cd33c442", "submitted_at": "2026-03-27T09:34:07Z", "updated_at": "2026-03-27T09:34:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.26211v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.26211", "identifiers": {"arxiv": "2603.26211"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.26211"}], "name": "Towards GUI Agents: Vision-Language Diffusion Models for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adapts a discrete vision-language diffusion model to single-turn GUI action and bounding-box prediction, using a hybrid masking schedule for coordinate geometry. Experiments across web, desktop and mobile datasets examine grounding accuracy and inference tradeoffs.", "topics": ["grounding", "action-prediction", "efficiency"], "updated_at": "2026-03-27", "url": "https://arxiv.org/abs/2603.26211"}
{"aliases": [], "authors": ["Rui Xie", "Zhi Gao", "Chenrui Shi", "Zirui Shang", "Lu Chen", "Qing Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b388086aeb8ac47e386099e5cd68cb89bb03978ee7d5897950181035d0abb723", "submitted_at": "2026-03-27T10:33:08Z", "updated_at": "2026-06-30T04:47:42Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.26266v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.26266", "identifiers": {"arxiv": "2603.26266"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.26266"}], "name": "GUIDE: Resolving Domain Bias in GUI Agents through Real-Time Web Video Retrieval and Plug-and-Play Annotation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-03-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUIDE retrieves relevant software tutorials through subtitle-based video search, then infers planning and grounding knowledge from consecutive annotated keyframes. It inserts that knowledge into existing GUI-agent modules without parameter updates and evaluates the plug-in on OSWorld.", "topics": ["video", "planning", "grounding", "training-free"], "updated_at": "2026-06-30", "url": "https://arxiv.org/abs/2603.26266"}
{"aliases": [], "authors": ["Zehai He", "Wenyi Hong", "Zhen Yang", "Ziyang Pan", "Mingdao Liu", "Xiaotao Gu", "Jie Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b2de420af47017d4c8aedd73ee07f749ab372e22dd7bac262bec684a68eb6880", "submitted_at": "2026-03-27T17:50:45Z", "updated_at": "2026-07-20T07:35:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.26648v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2603.26648", "identifiers": {"arxiv": "2603.26648"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.26648"}], "name": "Vision2Web: A Hierarchical Benchmark for Visual Website Development with Agent Verification", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-20", "url": "https://arxiv.org/abs/2603.26648"}
{"aliases": [], "authors": ["Zhaopeng Feng", "Liangcai Su", "Zhen Zhang", "Xinyu Wang", "Xiaotian Zhang", "Xiaobin Wang", "Runnan Fang", "Qi Zhang", "Baixuan Li", "Shihao Cai", "Rui Ye", "Hui Chen", "Jiang Yong", "Joey Tianyi Zhou", "Chenxiong Qian", "Pengjun Xie", "Bryan Hooi", "Zuozhu Liu", "Jingren Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "80db4d9d43ae6092071495d8a931bb888ae76226353906e9534cc7e5881e3685", "submitted_at": "2026-03-29T02:59:27Z", "updated_at": "2026-03-29T02:59:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.27490v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.27490", "identifiers": {"arxiv": "2603.27490"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.27490"}], "name": "AgentSwing: Adaptive Parallel Context Management Routing for Long-Horizon Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-29", "url": "https://arxiv.org/abs/2603.27490"}
{"aliases": [], "authors": ["Shi Qiu", "Junyi Deng", "Yiwei Deng", "Haoran Dong", "Jieyu Fu", "Mao Li", "Zeyu Li", "Zhaolong Zhang", "Huiwen Zheng", "Leidong Bao", "Anqi Lv", "Zihan Mo", "Yadi Niu", "Yiyang Peng", "Yu Tian", "Yili Wang", "Ziyu Wang", "Zi-Yu Wang", "Jiashen Wei", "Liuheng Wu", "Aoran Xue", "Leyi Yang", "Guanglu Yuan", "Xiarui Zhan", "Jingjun Zhang", "Zifan Zheng", "Pengfei Liu", "Linrui Zhen", "Kaiyang Li", "Qichang Li", "Ziheng Zhou", "Guo-En Nian", "Yunwei Xiao", "Qing-Hong Cao", "Linjie Dai", "Xu Feng", "Peng Gao", "Ying Gu", "Chang Liu", "Jia Liu", "Ming-xing Luo", "Yan-Qing Ma", "Liang-You Peng", "Huichao Song", "Shufeng Wang", "Chenxu Wang", "Tao Wang", "Yi-Nan Wang", "Chengyin Wu", "Pengwei Zhao", "Hua Xing Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29e37c2b42c6069d3a65c518d150c1a3d805bbbe367aad150191aaedca1c682d", "submitted_at": "2026-03-29T11:44:57Z", "updated_at": "2026-03-29T11:44:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.27646v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.27646", "identifiers": {"arxiv": "2603.27646"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.27646"}], "name": "PRBench: End-to-end Paper Reproduction in Physics Research", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-29", "url": "https://arxiv.org/abs/2603.27646"}
{"aliases": [], "authors": ["Christopher Clark", "Yue Yang", "Jae Sung Park", "Zixian Ma", "Jieyu Zhang", "Rohun Tripathi", "Mohammadreza Salehi", "Sangho Lee", "Taira Anderson", "Winson Han", "Ranjay Krishna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fe857c9f7e7819a9c2f99448f77bbeae15ae4e2ac8ad308000f3d217b0177fc", "submitted_at": "2026-03-30T06:15:06Z", "updated_at": "2026-03-30T06:15:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.28069v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.28069", "identifiers": {"arxiv": "2603.28069"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.28069"}], "name": "MolmoPoint: Better Pointing for VLMs with Grounding Tokens", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent general VLM pointing method with an explicit GUI-grounding experiment; no autonomous GUI-task execution is established.", "published_at": "2026-03-30", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "MolmoPoint replaces textual coordinate generation with pointing tokens that select visual regions and refine them through subpatches. The general image/video method includes GUI-pointing evaluation on ScreenSpotPro alongside other visual grounding tasks.", "topics": ["grounding", "screen-understanding", "video", "efficiency"], "updated_at": "2026-03-30", "url": "https://arxiv.org/abs/2603.28069"}
{"aliases": [], "authors": ["Fangda Ye", "Yuxin Hu", "Pengxiang Zhu", "Yibo Li", "Ziqi Jin", "Yao Xiao", "Yibo Wang", "Lei Wang", "Zhen Zhang", "Lu Wang", "Yue Deng", "Bin Wang", "Yifan Zhang", "Liangcai Su", "Xinyu Wang", "He Zhao", "Chen Wei", "Qiang Ren", "Bryan Hooi", "An Bo", "Shuicheng Yan", "Lidong Bing"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f3c628d03df650537e2484e3c7917fa07156691d9af5d8e6409e3321c4a85720", "submitted_at": "2026-03-30T13:16:03Z", "updated_at": "2026-03-30T13:16:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.28407v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.28407", "identifiers": {"arxiv": "2603.28407"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.28407"}], "name": "MiroEval: Benchmarking Multimodal Deep Research Agents in Process and Outcome", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-30", "url": "https://arxiv.org/abs/2603.28407"}
{"aliases": [], "authors": ["Zifan Peng", "Mingchen Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "31c9b9b8489ee3e241d86c191b9b3a17c756e292fae61cd09371ae11d04aef6b", "submitted_at": "2026-03-30T15:12:55Z", "updated_at": "2026-03-31T09:11:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.28551v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.28551", "identifiers": {"arxiv": "2603.28551"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.28551"}], "name": "\"What Did It Actually Do?\": Understanding Risk Awareness and Traceability for Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-31", "url": "https://arxiv.org/abs/2603.28551"}
{"aliases": [], "authors": ["Deepak Akkil", "Mowafak Allaham", "Amal Raj", "Tamer Abuelsaad", "Ravi Kokku"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4f03e65387e3aad99b8c4cef166a7a8b4a9e2954ad932b8daf566e30f156595", "submitted_at": "2026-03-30T21:27:28Z", "updated_at": "2026-03-30T21:27:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.29020v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.29020", "identifiers": {"arxiv": "2603.29020"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.29020"}], "name": "Emergence WebVoyager: Toward Consistent and Transparent Evaluation of (Web) Agents in The Wild", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-03-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Audits sources of ambiguity and operational variation in WebVoyager evaluation and introduces standardized guidance for task setup, failure handling, annotation and reporting. Emergence WebVoyager applies this protocol to a web-agent evaluation.", "topics": ["evaluation", "web-navigation"], "updated_at": "2026-03-30", "url": "https://arxiv.org/abs/2603.29020"}
{"aliases": [], "authors": ["Hongyi Nie", "Xunyuan Liu", "Yudong Bai", "Yaqing Wang", "Yang Liu", "Quanming Yao", "Zhen Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c71b713efd6ea47e7c78f5dd51d06c5cfdba750529bfbcdfc8e1c69897cbcc9", "submitted_at": "2026-03-31T06:37:12Z", "updated_at": "2026-03-31T06:37:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.29318v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2603.29318", "identifiers": {"arxiv": "2603.29318"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.29318"}], "name": "PSPA-Bench: A Personalized Benchmark for Smartphone GUI Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-03-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "PSPA-Bench evaluates personalized smartphone GUI behavior using user-specific instructions and a process-aware assessment of interaction traces. Its comparison of GUI agents diagnoses weaknesses in adapting workflows and preferences rather than measuring generic task completion alone.", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2026-03-31", "url": "https://arxiv.org/abs/2603.29318"}
{"aliases": [], "authors": ["Meiwen Ding", "Song Xia", "Chenqi Kong", "Xudong Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "894dc4ea8371b068c2ecccb79e282a392f8ab47c49c6f3420a1834a4e70a3592", "submitted_at": "2026-03-31T08:22:07Z", "updated_at": "2026-08-11T15:52:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.29418v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.29418", "identifiers": {"arxiv": "2603.29418"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.29418"}], "name": "Covert Visual Prompt Injection against Commercial Multimodal Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-11", "url": "https://arxiv.org/abs/2603.29418"}
{"aliases": [], "authors": ["Soslan Kabisov", "Vsevolod Kirichuk", "Andrey Volkov", "Gennadii Savrasov", "Marina Barannikov", "Anton Konushin", "Andrey Kuznetsov", "Dmitrii Zhemchuzhnikov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "31e0e9f980ecef87c239f0f019273551691d165b4cadb314947a7e832dfc1817", "submitted_at": "2026-02-18T10:27:29Z", "updated_at": "2026-10-06T20:55:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.29847v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2603.29847", "identifiers": {"arxiv": "2603.29847"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2603.29847"}], "name": "CADReasoner: Iterative Program Editing for CAD Reverse Engineering", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-06", "url": "https://arxiv.org/abs/2603.29847"}
{"aliases": [], "authors": ["Patrice Bechard", "Orlando Marquez Ayala", "Emily Chen", "Jordan Skelton", "Sagar Davasam", "Srinivas Sunkara", "Vikas Yadav", "Sai Rajeswar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "93eb667bf6af8fe98bdbf5930c6a2f67a7fa0b74dda8b3f0667ddefa58a0aa7b", "submitted_at": "2026-03-31T14:14:47Z", "updated_at": "2026-10-14T14:23:23Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2604.00073v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2604.00073", "identifiers": {"arxiv": "2604.00073"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.00073"}], "name": "Terminal Agents Suffice for Enterprise Automation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2604.00073"}
{"aliases": [], "authors": ["Haibo Wang", "Zihao Lin", "Zhiyang Xu", "Lifu Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "713f0e74d08e16d2cecd1f9233d88b6580b249682675a0069d6d64d047b61e94", "submitted_at": "2026-04-01T06:12:16Z", "updated_at": "2026-04-02T06:20:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.00528v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2604.00528", "identifiers": {"arxiv": "2604.00528"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.00528"}], "name": "Think, Act, Build: An Agentic Framework with Vision Language Models for Zero-Shot 3D Visual Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-02", "url": "https://arxiv.org/abs/2604.00528"}
{"aliases": [], "authors": ["Deepak Nathani", "Cheng Zhang", "Chang Huan", "Jiaming Shan", "Yinfei Yang", "Alkesh Patel", "Zhe Gan", "William Yang Wang", "Michael Saxon", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ae0f95c0c5ee3ca8a9eaf417143fb038f6f37f229a18eadffa8c1ade44a14c7", "submitted_at": "2026-04-01T12:53:01Z", "updated_at": "2026-04-01T12:53:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.00842v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "id": "arxiv:2604.00842", "identifiers": {"arxiv": "2604.00842"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.00842"}], "name": "Proactive Agent Research Environment: Simulating Active Users to Evaluate Proactive Assistants", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-01", "url": "https://arxiv.org/abs/2604.00842"}
{"aliases": [], "authors": ["Nan Wang", "Zhiwei Jin", "Chen Chen", "Haonan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce7f70ff1f8c7d4058d8fdf00cb6cf1d61c0367a50ba76fc96d356bb07452c60", "submitted_at": "2026-04-01T13:33:27Z", "updated_at": "2026-04-01T13:33:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.00886v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2604.00886", "identifiers": {"arxiv": "2604.00886"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.00886"}], "name": "PixelPrune: Pixel-Level Adaptive Visual Token Reduction via Predictive Coding", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent cross-domain VLM compression method with actual GUI-benchmark evaluation; the abstract does not establish a full computer-use policy.", "published_at": "2026-04-01", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "PixelPrune removes pixel-redundant patches before the vision encoder, reducing both encoding and downstream language-model work. The training-free method evaluates lossless and controlled-lossy compression on document and GUI tasks.", "topics": ["efficiency", "screen-understanding", "training-free", "evaluation"], "updated_at": "2026-04-01", "url": "https://arxiv.org/abs/2604.00886"}
{"aliases": [], "authors": ["Henry Peng Zou", "Chunyu Miao", "Wei-Chieh Huang", "Yankai Chen", "Yue Zhou", "Hanrong Zhang", "Yaozu Wu", "Liancheng Fang", "Zhengyao Gu", "Zhen Zhang", "Kening Zheng", "Fangxin Wang", "Yi Nian", "Shanghao Li", "Wenzhe Fan", "Langzhou He", "Weizhi Zhang", "Xue Liu", "Philip S. Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ad47ef265807999774de6f3098a7f8b52379543200a84b4a41e216e2c788b54c", "submitted_at": "2026-04-01T13:40:28Z", "updated_at": "2026-04-01T13:40:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.00892v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.00892", "identifiers": {"arxiv": "2604.00892"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.00892"}], "name": "When Users Change Their Mind: Evaluating Interruptible Agents in Long-Horizon Web Navigation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-01", "url": "https://arxiv.org/abs/2604.00892"}
{"aliases": [], "authors": ["Zhengyang Tang", "Ke Ji", "Xidong Wang", "Zihan Ye", "Xinyuan Wang", "Yiduo Guo", "Ziniu Li", "Chenxin Li", "Jingyuan Hu", "Shunian Chen", "Tongxu Luo", "Jiaxi Bi", "Zeyu Qin", "Shaobo Wang", "Xin Lai", "Pengyuan Lyu", "Junyi Li", "Can Xu", "Chengquan Zhang", "Han Hu", "Ming Yan", "Benyou Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f137028b7608598fb1885a0bf1330e00c9404851571f856665aae1af1d9d63c", "submitted_at": "2026-04-01T14:50:50Z", "updated_at": "2026-04-02T13:51:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.00986v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2604.00986", "identifiers": {"arxiv": "2604.00986"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.00986"}], "name": "Do Phone-Use Agents Respect Your Privacy?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-04-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MyPhoneBench to verify whether phone-use agents minimize personal data disclosure when filling forms during benign mobile tasks.", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2026-04-02", "url": "https://arxiv.org/abs/2604.00986"}
{"aliases": [], "authors": ["Yanting Wang", "Wei Zou", "Runpeng Geng", "Jinyuan Jia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "71d3120b6536126836038a2cfa3662896a9c08ab74adb866b5fa44ed2954a5da", "submitted_at": "2026-04-01T17:40:03Z", "updated_at": "2026-07-25T19:27:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.01194v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2604.01194", "identifiers": {"arxiv": "2604.01194"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.01194"}], "name": "AgentWatcher: A Rule-based Prompt Injection Monitor", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-25", "url": "https://arxiv.org/abs/2604.01194"}
{"aliases": [], "authors": ["Zhe Yang", "Shulin Tian", "Kairui Hu", "Shuai Liu", "Hoang-Nhat Nguyen", "Yichi Zhang", "Zujin Guo", "Mengying Yu", "Zinan Zhang", "Jingkang Yang", "Chen Change Loy", "Ziwei Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a3e9d52421c7f4bc23692a5f6f154919440600e30a30a44e336fde0502ffdab6", "submitted_at": "2026-04-01T17:58:33Z", "updated_at": "2026-04-01T17:58:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.01221v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.01221", "identifiers": {"arxiv": "2604.01221"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.01221"}], "name": "HippoCamp: Benchmarking Contextual Agents on Personal Computers", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-01", "url": "https://arxiv.org/abs/2604.01221"}
{"aliases": [], "authors": ["Bowen Wei", "Yunbei Zhang", "Jinhao Pan", "Kai Mei", "Xiao Wang", "Jihun Hamm", "Ziwei Zhu", "Yingqiang Ge"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "42c31b9e77a0908dd380e92890135c9d0d11558ca7818c908acea00d2d6c0045", "submitted_at": "2026-04-01T22:24:24Z", "updated_at": "2026-04-04T15:18:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.01438v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2604.01438", "identifiers": {"arxiv": "2604.01438"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.01438"}], "name": "ClawSafety: \"Safe\" LLMs, Unsafe Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-04", "url": "https://arxiv.org/abs/2604.01438"}
{"aliases": [], "authors": ["Masafumi Enomoto", "Ryoma Obara", "Haochen Zhang", "Masafumi Oyamada"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "58f09323010bc9e11bf87ca06cb5cd1d4d9602290a645f67a1e634d9fc45e3ae", "submitted_at": "2026-04-02T02:14:47Z", "updated_at": "2026-04-02T02:14:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.01535v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.01535", "identifiers": {"arxiv": "2604.01535"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.01535"}], "name": "Read More, Think More: Revisiting Observation Reduction for Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-02", "url": "https://arxiv.org/abs/2604.01535"}
{"aliases": [], "authors": ["Qi Zhang", "Shen Huang", "Chu Liu", "Shouqing Yang", "Junbo Zhao", "Haobo Wang", "Pengjun Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd063b58a0a5901825546bfd61d1c4c2b8a2bc1bffa811f5371b64e4af4b3d46", "submitted_at": "2026-04-02T03:13:53Z", "updated_at": "2026-04-02T03:13:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.01560v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2604.01560", "identifiers": {"arxiv": "2604.01560"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.01560"}], "name": "DeltaMem: Towards Agentic Memory Management via Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-02", "url": "https://arxiv.org/abs/2604.01560"}
{"aliases": [], "authors": ["Zirui Zhao", "Jun Hao Liew", "Yan Yang", "Wenzhuo Yang", "Ziyang Luo", "Doyen Sahoo", "Silvio Savarese", "Junnan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47e2cb6818c8e26889da590d26073ccc84474bb2e3391f9b184e5d7f8a31ef1d", "submitted_at": "2026-04-02T06:27:52Z", "updated_at": "2026-04-04T01:55:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.01676v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.01676", "identifiers": {"arxiv": "2604.01676"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.01676"}], "name": "GPA: Learning GUI Process Automation from Demonstrations", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents demonstration-based GUI process replay with Sequential Monte Carlo localization and readiness checks. GPA executes the learned workflow locally and can serve as the GUI execution component of an agent using MCP or CLI tools.", "topics": ["automation", "workflow", "robustness"], "updated_at": "2026-04-04", "url": "https://arxiv.org/abs/2604.01676"}
{"aliases": [], "authors": ["Mengzhou Wu", "Yuzhe Guo", "Yuan Cao", "Haochuan Lu", "Songhe Zhu", "Pingzhe Qu", "Xin Chen", "Kang Qin", "Zhongpu Wang", "Xiaode Zhang", "Xinyi Wang", "Wei Dai", "Gang Cao", "Yuetang Deng", "Zhi Gong", "Dezhi Ran", "Linyi Li", "Wei Yang", "Tao Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bac86c2ff2bf18feb3edfbe76285d33718d0c6d9564aff5a9d62f8778cdc5147", "submitted_at": "2026-02-11T17:02:38Z", "updated_at": "2026-02-11T17:02:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.02345v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.02345", "identifiers": {"arxiv": "2604.02345"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.02345"}], "name": "UI-Oceanus: Scaling GUI Agents with Synthetic Environmental Dynamics", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-02-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-Oceanus converts autonomous GUI exploration and execution feedback into supervision for predicting future interface states. It compares forward and inverse objectives during continual pretraining and examines downstream offline and online navigation as synthetic data scales.", "topics": ["world-models", "data-generation", "state-tracking", "evaluation"], "updated_at": "2026-02-11", "url": "https://arxiv.org/abs/2604.02345"}
{"aliases": [], "authors": ["Wei Zou", "Mingwen Dong", "Miguel Romero Calvo", "Shuaichen Chang", "Jiang Guo", "Dongkyu Lee", "Xing Niu", "Xiaofei Ma", "Yanjun Qi", "Jiarong Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a37f4bdc495e10e4ae301ebff6faa28cdfa6de152d4619c051083b5dce41e6ba", "submitted_at": "2026-04-03T01:25:12Z", "updated_at": "2026-04-07T14:45:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.02623v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.02623", "identifiers": {"arxiv": "2604.02623"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.02623"}], "name": "Poison Once, Exploit Forever: Environment-Injected Memory Poisoning Attacks on Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies persistent web-agent memory contamination introduced through environmental observations instead of direct memory access. Experiments on WebArena and VisualWebArena examine cross-session effects and how interaction failures alter susceptibility.", "topics": ["safety-security", "memory", "web-navigation"], "updated_at": "2026-04-07", "url": "https://arxiv.org/abs/2604.02623"}
{"aliases": [], "authors": ["Yunhao Feng", "Yifan Ding", "Yingshui Tan", "Xingjun Ma", "Yige Li", "Yutao Wu", "Yifeng Gao", "Kun Zhai", "Yanming Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "04d7feef296b78f107d7a718f8784962359a29a44569c4f128d89ad5640c736b", "submitted_at": "2026-04-03T10:29:31Z", "updated_at": "2026-09-22T02:57:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.02947v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.02947", "identifiers": {"arxiv": "2604.02947"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.02947"}], "name": "AgentHazard: A Benchmark for Evaluating Harmful Behavior in Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-22", "url": "https://arxiv.org/abs/2604.02947"}
{"aliases": [], "authors": ["Cornelius Kummer", "Lena Jurkschat", "Michael Färber", "Sahar Vahdati"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "34ca94c6b47a63c68383b72db7aa6fe1476566478263654d8854f0f8d2022d86", "submitted_at": "2026-04-03T11:41:53Z", "updated_at": "2026-04-03T11:41:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.02985v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "id": "arxiv:2604.02985", "identifiers": {"arxiv": "2604.02985"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.02985"}], "name": "Prompt Compression in the Wild: Measuring Latency, Rate Adherence, and Quality for Faster LLM Inference", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-03", "url": "https://arxiv.org/abs/2604.02985"}
{"aliases": [], "authors": ["Renze Lou", "Baolin Peng", "Wenlin Yao", "Qianhui Wu", "Hao Cheng", "Suman Nath", "Wenpeng Yin", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "959cf0d79bfcbf984b7aa1ea29359d7b25070fac07d30284caf7a86038efd715", "submitted_at": "2026-04-03T21:18:26Z", "updated_at": "2026-07-16T12:25:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.03465v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.03465", "identifiers": {"arxiv": "2604.03465"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.03465"}], "name": "The Tool Illusion: Rethinking Tool Use in Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-16", "url": "https://arxiv.org/abs/2604.03465"}
{"aliases": [], "authors": ["Jackson Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8da5e35fb4244b6fbd9aed8424d5f417539de91a64a9a36890f4d8bf48c45184", "submitted_at": "2026-04-04T05:49:30Z", "updated_at": "2026-04-04T05:49:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.03598v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.03598", "identifiers": {"arxiv": "2604.03598"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.03598"}], "name": "AttackEval: A Systematic Empirical Study of Prompt Injection Attack Effectiveness Against Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-04", "url": "https://arxiv.org/abs/2604.03598"}
{"aliases": [], "authors": ["Yuwen Zhai", "Runze Li", "Liang Wang", "Nian Shi", "Liwu Xu", "Wei Zhang", "Ran Lin", "Bo Xu", "Benlei Cui"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "856e9920aea7d821061b3258b31aa32fd37b5579d628f5507f45ba83f237b77b", "submitted_at": "2026-04-06T03:58:43Z", "updated_at": "2026-04-06T03:58:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.04399v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.04399", "identifiers": {"arxiv": "2604.04399"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.04399"}], "name": "GUIDE: Interpretable GUI Agent Evaluation via Hierarchical Diagnosis", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUIDE evaluates GUI trajectories by segmenting them into subtasks, diagnosing each segment, and aggregating structured findings into a task verdict. It tests diagnostic judgments on e-commerce, web-agent and Android-control traces rather than relying on one holistic trajectory score.", "topics": ["evaluation", "long-horizon", "state-tracking"], "updated_at": "2026-04-06", "url": "https://arxiv.org/abs/2604.04399"}
{"aliases": [], "authors": ["Jingyang Qiao", "Weicheng Meng", "Yu Cheng", "Zhihang Lin", "Zhizhong Zhang", "Xin Tan", "Jingyu Gong", "Kun Shao", "Yuan Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4a94c25fd6cca29a38820023d4084b557d7a920f9a0e071e5aa6b1a0373fb14e", "submitted_at": "2026-04-06T07:59:52Z", "updated_at": "2026-04-19T09:08:36Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2604.04503v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2604.04503", "identifiers": {"arxiv": "2604.04503"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.04503"}], "name": "Memory Intelligence Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-19", "url": "https://arxiv.org/abs/2604.04503"}
{"aliases": [], "authors": ["Zijun Wang", "Haoqin Tu", "Letian Zhang", "Hardy Chen", "Juncheng Wu", "Xiangyan Liu", "Zhenlong Yuan", "Tianyu Pang", "Michael Qizhe Shieh", "Fengze Liu", "Zeyu Zheng", "Huaxiu Yao", "Yuyin Zhou", "Cihang Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ee3cdf381ded2dff50d0f672f2eba49206d90b712477016d889b4414753786c4", "submitted_at": "2026-04-06T15:27:05Z", "updated_at": "2026-04-06T15:27:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.04759v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.04759", "identifiers": {"arxiv": "2604.04759"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.04759"}], "name": "Your Agent, Their Asset: A Real-World Safety Analysis of OpenClaw", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-06", "url": "https://arxiv.org/abs/2604.04759"}
{"aliases": [], "authors": ["Chaoran Chen", "Zhiping Zhang", "Zeya Chen", "Eryue Xu", "Yinuo Yang", "Ibrahim Khalilov", "Simret A Gebreegziabher", "Yanfang Ye", "Ziang Xiao", "Yaxing Yao", "Tianshi Li", "Toby Jia-Jun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1411ee71ad8f65e4941af4ef538f52da6e9bf9c57fb537b6ecf37c8478609693", "submitted_at": "2026-04-06T17:57:16Z", "updated_at": "2026-09-12T06:05:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.04918v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.04918", "identifiers": {"arxiv": "2604.04918"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.04918"}], "name": "Comparing Human Oversight Strategies for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares four human oversight strategies during live web-agent sessions, examining exposure to problematic actions and opportunities to intervene. The analysis identifies timing and attention problems in control handoffs and upfront plan approval.", "topics": ["human-in-the-loop", "safety-security", "evaluation"], "updated_at": "2026-09-12", "url": "https://arxiv.org/abs/2604.04918"}
{"aliases": [], "authors": ["Rongqian Chen", "Yu Li", "Zeyu Fang", "Sizhe Tang", "Weidong Cao", "Tian Lan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "238aac264fb21022f7c546ae3aa3f4ea104c381f5d318f6d962f22fb00095216", "submitted_at": "2026-04-06T20:39:30Z", "updated_at": "2026-05-28T08:13:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.05157v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.05157", "identifiers": {"arxiv": "2604.05157"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.05157"}], "name": "IntentScore: Intent-Conditioned Action Evaluation for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains IntentScore on offline desktop GUI interactions to assess candidate actions using both their state relevance and planning intent. The reward model is evaluated as an action reranker for Agent S3 on OSWorld.", "topics": ["reward-models", "planning", "evaluation"], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2604.05157"}
{"aliases": [], "authors": ["Yuzhe Zhang", "Xianwei Xue", "Xingyong Wu", "Mengke Chen", "Chen Liu", "Xinran He", "Run Shao", "Feiran Liu", "Huanmin Xu", "Qiutong Pan", "Haiwei Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c9ef656cbee786bec0d37ac4915bfd04adb36410eda8e647fab0d8605f90753", "submitted_at": "2026-04-07T06:16:23Z", "updated_at": "2026-04-07T06:16:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.05477v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.05477", "identifiers": {"arxiv": "2604.05477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.05477"}], "name": "Don't Act Blindly: Robust GUI Automation via Action-Effect Verification and Self-Correction", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces VeriGUI, which checks expected action effects and reasons about recovery before subsequent GUI actions. Training combines synthetic failure trajectories with verification-weighted GRPO, evaluated on an AndroidControl-based robustness benchmark.", "topics": ["robustness", "reinforcement-learning", "state-tracking"], "updated_at": "2026-04-07", "url": "https://arxiv.org/abs/2604.05477"}
{"aliases": [], "authors": ["Liyuan Deng", "Shujian Deng", "Yongkang Chen", "Yongkang Dai", "Zhihang Zhong", "Linyang Li", "Xiao Sun", "Yilei Shi", "Huaxi Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1d0dddc7f572a48f1b615e60040e90dc7b185de63382841edf01bb02d9ac243b", "submitted_at": "2026-04-07T07:45:17Z", "updated_at": "2026-04-07T07:45:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.05547v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.05547", "identifiers": {"arxiv": "2604.05547"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.05547"}], "name": "COSMO-Agent: Tool-Augmented Agent for Closed-loop Optimization,Simulation,and Modeling Orchestration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-07", "url": "https://arxiv.org/abs/2604.05547"}
{"aliases": [], "authors": ["Michael Krumdick", "Varshini Reddy", "Shivani Chaudhary", "William Day", "Maarij Ahmed", "Hayan Haqqi", "Muhammad Ahsen Fahim", "Hanzallah Amjad", "Ahmad Orakzai", "Aqsa Gul", "Chris Tanner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "79f2481866d3b3cf067447bf4a198027ffd41a7e9e77c8a17ea21f58937f2dce", "submitted_at": "2026-04-07T14:15:45Z", "updated_at": "2026-04-07T14:15:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.05912v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.05912", "identifiers": {"arxiv": "2604.05912"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.05912"}], "name": "FrontierFinance: A Long-Horizon Computer-Use Benchmark of Real-World Financial Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-07", "url": "https://arxiv.org/abs/2604.05912"}
{"aliases": [], "authors": ["Pranjal Aggarwal", "Graham Neubig", "Sean Welleck"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fdd9cb6daad8832261ebe861fc1cb74429b819305e14ecfeb9698616013e6ac8", "submitted_at": "2026-04-07T17:38:15Z", "updated_at": "2026-04-07T17:38:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.06126v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.06126", "identifiers": {"arxiv": "2604.06126"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.06126"}], "name": "Gym-Anything: Turn any Software into an Agent Environment", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Gym-Anything, where a setup agent configures software environments and an independent auditor checks the resulting evidence. CUA-World supplies long-horizon tasks across professional applications, with experiments on trajectory distillation and test-time auditing.", "topics": ["data-generation", "sandbox", "long-horizon", "evaluation"], "updated_at": "2026-04-07", "url": "https://arxiv.org/abs/2604.06126"}
{"aliases": [], "authors": ["Sangwook Lee", "Adnan Abbas", "Sang Won Lee", "Young-Ho Kim", "Yan Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "104f2e3f9f9d3c69b9443a90c6de7d33476a3c8155f464b486c0b0ba00ec95eb", "submitted_at": "2026-04-07T17:44:19Z", "updated_at": "2026-08-09T18:02:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.06134v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.06134", "identifiers": {"arxiv": "2604.06134"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.06134"}], "name": "MAESTRO: Adapting GUIs and Guiding Navigation with User Preferences in Conversational Agents with GUIs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-09", "url": "https://arxiv.org/abs/2604.06134"}
{"aliases": [], "authors": ["Komal Kumar", "Aman Chadha", "Salman Khan", "Fahad Shahbaz Khan", "Hisham Cholakkal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "17ffc141358c28b24e6de27817b4c0f9969f58ba9ea09b6586ad0c9dd4094f27", "submitted_at": "2026-04-07T17:59:58Z", "updated_at": "2026-04-07T17:59:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.06170v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.06170", "identifiers": {"arxiv": "2604.06170"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.06170"}], "name": "Paper Circle: An Open-source Multi-agent Research Discovery and Analysis Framework", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-07", "url": "https://arxiv.org/abs/2604.06170"}
{"aliases": [], "authors": ["Yuelin Hu", "Zhengxue Cheng", "Ronghua Wu", "Qunshan Gu", "Hongwei Hu", "Wei Liu", "Qiao Liang", "Li Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f051480f6f95d0c8b8a387d780a0de138beca556cb93dca470ed8818cdaf0d5e", "submitted_at": "2026-02-03T08:13:05Z", "updated_at": "2026-02-03T08:13:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.06177v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.06177", "identifiers": {"arxiv": "2604.06177"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.06177"}], "name": "WebExpert: domain-aware web agents with critic-guided expert experience for high-precision search", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-03", "url": "https://arxiv.org/abs/2604.06177"}
{"aliases": [], "authors": ["Yichen Gong", "Zhuohan Cai", "Sunhao Dai", "Yuqi Zhou", "Zhangxuan Gu", "Changhua Meng", "Shuheng Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8790de644f655f4e2a27fa300e9e9e07de1071fa5b50bf3e7a0945c9989eacfc", "submitted_at": "2026-02-06T18:54:42Z", "updated_at": "2026-02-06T18:54:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.06182v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.06182", "identifiers": {"arxiv": "2604.06182"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.06182"}], "name": "VenusBench-Mobile: A Challenging and User-Centric Benchmark for Mobile GUI Agents with Capability Diagnostics", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-06", "url": "https://arxiv.org/abs/2604.06182"}
{"aliases": [], "authors": ["Corby Rosset", "Pratyusha Sharma", "Andrew Zhao", "Miguel Gonzalez-Fernandez", "Ahmed Awadallah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b452684b305c4f20ded686289b448fbe4f08bb3b34f515a43e2ce5d6dc4ef506", "submitted_at": "2026-04-05T06:26:05Z", "updated_at": "2026-04-05T06:26:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.06240v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.06240", "identifiers": {"arxiv": "2604.06240"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.06240"}], "name": "The Art of Building Verifiers for Computer Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a web-trajectory verifier built around structured rubrics, separate process and outcome judgments, and explicit treatment of uncontrollable failures. CUAVerifierBench supplies human labels for evaluating these choices and screenshot-based context management.", "topics": ["reward-models", "evaluation", "long-horizon"], "updated_at": "2026-04-05", "url": "https://arxiv.org/abs/2604.06240"}
{"aliases": [], "authors": ["Guruprasad Viswanathan Ramesh", "Asmit Nayak", "Basieem Siddique", "Kassem Fawaz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "400d8a9f7aed4f150dd5198d9a1561d2df0454069495594ca165ad4254a1be59", "submitted_at": "2026-04-07T18:43:21Z", "updated_at": "2026-06-16T22:37:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.06367v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.06367", "identifiers": {"arxiv": "2604.06367"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.06367"}], "name": "WebSP-Eval: Evaluating Web Agents on Website Security and Privacy Tasks", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-04-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebSP-Eval evaluates agents carrying out website security and privacy tasks such as preference changes and session management. A browser extension controls account and initial state across runs, enabling analysis of failures on stateful UI controls.", "topics": ["safety-security", "evaluation", "web-navigation", "state-tracking"], "updated_at": "2026-06-16", "url": "https://arxiv.org/abs/2604.06367"}
{"aliases": [], "authors": ["Songze Li", "Xiaoke Guo", "Tianqi Liu", "Biao Yi", "Zhaoyan Gong", "Zhiqiang Liu", "Huajun Chen", "Wen Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b99643e43c5e20af088d6fa62152e03256c947af76a6b3c0abfc6f74a0f7d648", "submitted_at": "2026-04-08T12:12:09Z", "updated_at": "2026-10-02T06:30:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.06995v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.06995", "identifiers": {"arxiv": "2604.06995"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.06995"}], "name": "What's Missing in Screen-to-Action? Towards a UI-in-the-Loop Paradigm for Multimodal GUI Reasoning", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces UILoop, which inserts explicit UI-element localization, function and usage reasoning between screen observation and action. UI Comprehension-Bench evaluates these element-level capabilities with three metrics.", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": "2026-10-02", "url": "https://arxiv.org/abs/2604.06995"}
{"aliases": [], "authors": ["Guo Gan", "Yuxuan Ding", "Cong Chen", "Yuwei Ren", "Yin Huang", "Hong Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f4d0dbeac950760f4f676ab7c651a881fd275f4adb69d46a94a11c4f10a03980", "submitted_at": "2026-04-08T16:40:26Z", "updated_at": "2026-05-08T03:06:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.07277v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.07277", "identifiers": {"arxiv": "2604.07277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.07277"}], "name": "Android Coach: Improve Online Agentic Training Efficiency with Single State Multiple Actions", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Android Coach, which uses a learned critic to evaluate multiple candidate actions at each observed emulator state. Process rewards and a group-wise advantage estimator support online training, evaluated on AndroidLab and AndroidWorld.", "topics": ["reinforcement-learning", "reward-models", "efficiency"], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2604.07277"}
{"aliases": [], "authors": ["Teng Li", "Ziyuan Huang", "Cong Chen", "Yangfu Li", "Yuanhuiyi Lyu", "Dandan Zheng", "Chunhua Shen", "Jun Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb80d6138d46f2fd4036e46fa8589ef7e59ef5e8bd9a42f619357e07e420017d", "submitted_at": "2026-04-08T17:53:52Z", "updated_at": "2026-04-08T17:53:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.07340v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.07340", "identifiers": {"arxiv": "2604.07340"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.07340"}], "name": "TC-AE: Unlocking Token Capacity for Deep Compression Autoencoders", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-08", "url": "https://arxiv.org/abs/2604.07340"}
{"aliases": [], "authors": ["Mingyu Ouyang", "Siyuan Hu", "Kevin Qinghong Lin", "Hwee Tou Ng", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc6b06c3b6c5bfad9b0ffc9dea79c7b623f03bb6be7538d94fddb334f68c63af", "submitted_at": "2026-04-08T17:49:03Z", "updated_at": "2026-04-08T17:49:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.07429v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.07429", "identifiers": {"arxiv": "2604.07429"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.07429"}], "name": "GameWorld: Towards Standardized and Verifiable Evaluation of Multimodal Game Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-08", "url": "https://arxiv.org/abs/2604.07429"}
{"aliases": [], "authors": ["Xing Han Lù", "Siva Reddy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "46d19668b7a5718e388e2dc00aa99025817b8287f6815bc7db4378179a271853", "submitted_at": "2026-04-09T04:04:15Z", "updated_at": "2026-04-09T04:04:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.07776v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.07776", "identifiers": {"arxiv": "2604.07776"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.07776"}], "name": "Structured Distillation of Web Agent Capabilities Enables Generalization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-09", "url": "https://arxiv.org/abs/2604.07776"}
{"aliases": [], "authors": ["Wenkui Yang", "Chao Jin", "Haisu Zhu", "Weilin Luo", "Derek Yuen", "Kun Shao", "Junxian Duan", "Huaibo Huang", "Jie Cao", "Ran He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5427eedff0e97f9c73101403754c38e8928ad71abab7b7eff5f136a30ee2205", "submitted_at": "2026-04-09T05:32:34Z", "updated_at": "2026-07-08T02:39:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.07831v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.07831", "identifiers": {"arxiv": "2604.07831"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.07831"}], "name": "Are GUI Agents Focused Enough? Automated Distraction via Semantic-level UI Element Injection", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-04-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study tests black-box GUI-grounding failures caused by strategically added, innocuous-looking UI elements. Its iterative overlay search is compared with random injection and repeated independent grounding trials to distinguish persistent misdirection from incidental clutter.", "topics": ["safety-security", "grounding", "robustness", "evaluation"], "updated_at": "2026-07-08", "url": "https://arxiv.org/abs/2604.07831"}
{"aliases": [], "authors": ["Maria Movin", "Claudia Hauff", "Aron Henriksson", "Panagiotis Papapetrou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f60f69f90ea1bec03745d3f72acea4a1f7b9810e0b46adaba6829a0d38cf626", "submitted_at": "2026-04-09T07:49:02Z", "updated_at": "2026-04-09T07:49:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.07929v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.07929", "identifiers": {"arxiv": "2604.07929"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.07929"}], "name": "Same Outcomes, Different Journeys: A Trace-Level Framework for Comparing Human and GUI-Agent Behavior in Production Search Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-09", "url": "https://arxiv.org/abs/2604.07929"}
{"aliases": [], "authors": ["Dominik Seip", "Matthias Hein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b97eed6788278e23f3aca57d11357c169d25396e90617ddefbf122b60a92e715", "submitted_at": "2026-04-09T09:08:40Z", "updated_at": "2026-08-10T12:07:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.08005v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.08005", "identifiers": {"arxiv": "2604.08005"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.08005"}], "name": "Preference Redirection via Attention Concentration: An Attack on Computer Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-10", "url": "https://arxiv.org/abs/2604.08005"}
{"aliases": [], "authors": ["Tongbo Chen", "Zhengxi Lu", "Zhan Xu", "Guocheng Shao", "Shaohan Zhao", "Fei Tang", "Yong Du", "Kaitao Song", "Yizhou Liu", "Yuchen Yan", "Wenqi Zhang", "Xu Tan", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e2954b76a0c92eb40f0bdff968c3fb746a909c851ca3c477fb62b0723742efa", "submitted_at": "2026-04-09T16:50:50Z", "updated_at": "2026-04-09T16:50:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.08455v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.08455", "identifiers": {"arxiv": "2604.08455"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.08455"}], "name": "KnowU-Bench: Towards Interactive, Proactive, and Personalized Mobile Agent Evaluation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-09", "url": "https://arxiv.org/abs/2604.08455"}
{"aliases": [], "authors": ["Tanmay Gupta", "Piper Wolters", "Zixian Ma", "Peter Sushko", "Rock Yuren Pang", "Diego Llanes", "Yue Yang", "Taira Anderson", "Boyuan Zheng", "Zhongzheng Ren", "Harsh Trivedi", "Taylor Blanton", "Caleb Ouellette", "Winson Han", "Ali Farhadi", "Ranjay Krishna"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dbc0d512a11b88a9036d06d0fcc26a8ab0d0bc54fad986d4b6225c8073897bc2", "submitted_at": "2026-04-09T17:54:02Z", "updated_at": "2026-04-09T17:54:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.08516v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.08516", "identifiers": {"arxiv": "2604.08516"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.08516"}], "name": "MolmoWeb: Open Visual Web Agent and Open Data for the Open Web", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The abstract states that artifacts will be released; this annotation does not independently verify release status or licenses.", "published_at": "2026-04-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents MolmoWebMix, combining synthetic browser trajectories, human demonstrations and GUI perception data, and MolmoWeb screenshot-to-action policies trained from it. Browser benchmarks also test parallel rollouts with best-of-N selection.", "topics": ["web-navigation", "data-generation", "imitation-learning"], "updated_at": "2026-04-09", "url": "https://arxiv.org/abs/2604.08516"}
{"aliases": [], "authors": ["Yuxuan Zhang", "Yubo Wang", "Yipeng Zhu", "Penghui Du", "Junwen Miao", "Xuan Lu", "Zhuofeng Li", "Xingwei Qu", "Zhengkang Guo", "Yuanzhe Shen", "Dingjie Song", "Han Zhou", "Tuney Zheng", "Xian Wu", "Hao Yu", "Songcheng Cai", "Yi Lu", "Yunzhuo Hao", "Minyi Lei", "Liang Chen", "Kai Zou", "Huifeng Yin", "Wendong Xu", "Dongfu Jiang", "Ping Nie", "Jiaheng Liu", "Wenhu Chen", "Kelsey R. Allen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "07c1dc2c31d2012b790a3497b5dfab558e8a663f92064c9f6996a770d3f5b8ad", "submitted_at": "2026-04-09T17:57:13Z", "updated_at": "2026-07-20T17:58:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.08523v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.08523", "identifiers": {"arxiv": "2604.08523"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.08523"}], "name": "ClawBench: Can AI Agents Complete Everyday Online Tasks?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-20", "url": "https://arxiv.org/abs/2604.08523"}
{"aliases": [], "authors": ["Yushi Feng", "Junye Du", "Qifan Wang", "Zizhan Ma", "Qian Niu", "Yutaka Matsuo", "Long Feng", "Lequan Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e53647528b9ee9d98f93ee665d0d886f9b400a129873e240704041df291f5749", "submitted_at": "2026-04-10T09:41:21Z", "updated_at": "2026-04-10T09:41:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09155v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.09155", "identifiers": {"arxiv": "2604.09155"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.09155"}], "name": "CORA: Conformal Risk-Controlled Agents for Safeguarded Mobile GUI Automation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The annotation describes risk calibration without asserting an unconditional safety guarantee.", "published_at": "2026-04-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CORA, which calibrates a risk model to choose between executing and abstaining from mobile GUI actions, then diagnoses rejected actions to suggest interventions. Phone-Harm provides step-level safety labels for evaluating this safeguarding process.", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2026-04-10", "url": "https://arxiv.org/abs/2604.09155"}
{"aliases": [], "authors": ["Jiachen Zhu", "Lingyu Yang", "Rong Shan", "Congmin Zheng", "Zeyu Zheng", "Weiwen Liu", "Yong Yu", "Weinan Zhang", "Jianghao Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d756e951c214d59e782faad5f12b784b2bea431a2ca08ae907f68dc07d61b50f", "submitted_at": "2026-02-24T04:29:42Z", "updated_at": "2026-02-24T04:29:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09574v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.09574", "identifiers": {"arxiv": "2604.09574"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.09574"}], "name": "Turing Test on Screen: A Benchmark for Mobile GUI Agent Humanization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-24", "url": "https://arxiv.org/abs/2604.09574"}
{"aliases": [], "authors": ["Wee Joe Tan", "Zi Rui Lucas Lim", "Shashank Durgad", "Karim Obegi", "Aiden Yiliu Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "445e520a60e0e3cc0c3c519e21bfa8eebfa33d4fe4320c437e67b94a02d84504", "submitted_at": "2026-02-25T18:59:42Z", "updated_at": "2026-04-14T20:54:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.09581v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.09581", "identifiers": {"arxiv": "2604.09581"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.09581"}], "name": "Avenir-UX: Automated UX Evaluation via Simulated Human Web Interaction with GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-02-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Avenir-UX combines grounded web interaction with simulated user profiles and structured usability measures to produce evaluations of website journeys. The paper describes its architecture and use of SUS, step-wise ease questions and think-aloud traces for automated UX reporting.", "topics": ["automation", "evaluation", "ui-design", "grounding"], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2604.09581"}
{"aliases": [], "authors": ["Yunfei Feng", "Xi Zhao", "Cheng Zhang", "Dahu Feng", "Daolin Cheng", "Jianqi Yu", "Yubin Xia", "Erhu Feng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "564a51339bcf082a06f42ead8eb1b475f9cc267773f2703084c502c084983c4d", "submitted_at": "2026-02-28T14:30:33Z", "updated_at": "2026-02-28T14:30:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09587v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.09587", "identifiers": {"arxiv": "2604.09587"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.09587"}], "name": "MobiFlow: Real-World Mobile Agent Benchmarking through Trajectory Fusion", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-02-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-02-28", "url": "https://arxiv.org/abs/2604.09587"}
{"aliases": [], "authors": ["Tiantian He", "Yihang Chen", "Keyue Jiang", "Ka Yiu Lee", "Kaiwen Zhou", "Kun Shao", "Shuai Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd2bc0ee1955dbc8e98a9306a3d6260251cbab89d36ac117e4dc49b6fa99766e", "submitted_at": "2026-04-10T18:46:05Z", "updated_at": "2026-04-10T18:46:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09815v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.09815", "identifiers": {"arxiv": "2604.09815"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.09815"}], "name": "EE-MCP: Self-Evolving MCP-GUI Agents via Automated Environment Generation and Experience Learning", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "EE-MCP studies application-dependent mixtures of GUI actions and MCP calls using generated environments, trajectory collection and filtered training. An experience bank extracts rules from trajectory comparisons for inference-time improvement, distinct from model distillation.", "topics": ["tool-integration", "data-generation", "imitation-learning", "memory"], "updated_at": "2026-04-10", "url": "https://arxiv.org/abs/2604.09815"}
{"aliases": [], "authors": ["Suhana Bedi", "Ryan Welch", "Ethan Steinberg", "Michael Wornow", "Taeil Matthew Kim", "Haroun Ahmed", "Peter Sterling", "Bravim Purohit", "Qurat Akram", "Angelic Acosta", "Esther Nubla", "Pritika Sharma", "Michael A. Pfeffer", "Sanmi Koyejo", "Nigam H. Shah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "55cedd4ad9760618fe9b3354e384f69b69b23b65f75837a3898484c9c6e55586", "submitted_at": "2026-04-10T22:33:39Z", "updated_at": "2026-04-10T22:33:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09937v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.09937", "identifiers": {"arxiv": "2604.09937"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.09937"}], "name": "HealthAdminBench: Evaluating Computer-Use Agents on Healthcare Administration Tasks", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces HealthAdminBench with GUI environments for an electronic health record, payer portals and fax handling. Expert-defined administrative tasks are decomposed into verifiable subtasks to compare local progress with end-to-end workflow completion.", "topics": ["evaluation", "workflow", "industry"], "updated_at": "2026-04-10", "url": "https://arxiv.org/abs/2604.09937"}
{"aliases": [], "authors": ["Zae Myung Kim", "Dongseok Lee", "Jaehyung Kim", "Vipul Raheja", "Dongyeop Kang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebd6aa6ffaabd40e023f4b69cdacbb59b108ef2a79b5a184b99ec4d3676fc12d", "submitted_at": "2026-04-11T15:58:29Z", "updated_at": "2026-04-17T01:52:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.10261v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.10261", "identifiers": {"arxiv": "2604.10261"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.10261"}], "name": "The Amazing Agent Race: Strong Tool Users, Weak Navigators", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-17", "url": "https://arxiv.org/abs/2604.10261"}
{"aliases": [], "authors": ["Wanyi Chen", "Xiao Yang", "Xu Yang", "Tianming Sha", "Qizheng Li", "Zhuo Wang", "Bowen Xian", "Fang Kong", "Weiqing Liu", "Jiang Bian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "070309fb07edfcc17c57feff0115e9268b6cfea60ec1f6f19b959be975170b88", "submitted_at": "2026-04-12T09:35:27Z", "updated_at": "2026-05-13T11:50:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.10547v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.10547", "identifiers": {"arxiv": "2604.10547"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.10547"}], "name": "Agent^2 RL-Bench: Can LLM Agents Engineer Agentic RL Post-Training?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-13", "url": "https://arxiv.org/abs/2604.10547"}
{"aliases": [], "authors": ["Xuwei Ding", "Skylar Zhai", "Linxin Song", "Jiate Li", "Taiwei Shi", "Nicholas Meade", "Siva Reddy", "Jian Kang", "Jieyu Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cffb933e5ed31a9cdfb7d644b2ceecb8114630061b85860eb890cebdcd2d0a92", "submitted_at": "2026-04-12T10:52:42Z", "updated_at": "2026-04-17T07:20:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.10577v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.10577", "identifiers": {"arxiv": "2604.10577"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.10577"}], "name": "The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-17", "url": "https://arxiv.org/abs/2604.10577"}
{"aliases": [], "authors": ["Shanshan Zhong", "Kate Shen", "Chenyan Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5bd703e000225c33d850eb2cd99a9fecde010b24e21ee33a99b6b973e2f8ffcb", "submitted_at": "2026-04-13T03:18:34Z", "updated_at": "2026-04-13T03:18:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.10938v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.10938", "identifiers": {"arxiv": "2604.10938"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.10938"}], "name": "AgentWebBench: Benchmarking Multi-Agent Coordination in Agentic Web", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-13", "url": "https://arxiv.org/abs/2604.10938"}
{"aliases": [], "authors": ["Peng Yuan", "Yuyang Yin", "Yuxuan Cai", "Zheng Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1db8d789fb8ba2d5e61ca24536670fd83491c305f532adfecf7193821f7eea52", "submitted_at": "2026-04-13T04:45:27Z", "updated_at": "2026-04-13T04:45:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.10988v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.10988", "identifiers": {"arxiv": "2604.10988"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.10988"}], "name": "WebForge: Breaking the Realism-Reproducibility-Scalability Trilemma in Browser Agent Benchmark", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents WebForge, a pipeline that plans, generates, refines and validates self-contained interactive websites. WebForge-Bench uses these environments to evaluate browser agents across domains and controlled difficulty levels.", "topics": ["data-generation", "evaluation", "web-navigation"], "updated_at": "2026-04-13", "url": "https://arxiv.org/abs/2604.10988"}
{"aliases": [], "authors": ["CocoaBench Team", "Shibo Hao", "Zhining Zhang", "Zhiqi Liang", "Tianyang Liu", "Yuheng Zha", "Qiyue Gao", "Jixuan Chen", "Zilong Wang", "Zhoujun Cheng", "Haoxiang Zhang", "Junli Wang", "Hexi Jin", "Boyuan Zheng", "Kun Zhou", "Yu Wang", "Feng Yao", "Licheng Liu", "Yijiang Li", "Zhifei Li", "Zhengtao Han", "Pracha Promthaw", "Tommaso Cerruti", "Xiaohan Fu", "Ziqiao Ma", "Jingbo Shang", "Lianhui Qin", "Julian McAuley", "Eric P. Xing", "Zhengzhong Liu", "Rupesh Kumar Srivastava", "Zhiting Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d7b9b6db5cf0d5b43f8021028c2a4d045dd4be7a639eac8f70a9e46f69f1a9e", "submitted_at": "2026-04-13T09:00:10Z", "updated_at": "2026-04-14T08:14:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.11201v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.11201", "identifiers": {"arxiv": "2604.11201"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.11201"}], "name": "CocoaBench: Evaluating Unified Digital Agents in the Wild", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2604.11201"}
{"aliases": [], "authors": ["Zhixin Lin", "Jungang Li", "Dongliang Xu", "Shidong Pan", "Yibo Shi", "Yuchi Liu", "Yuecong Min", "Yue Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df9bec627d3ffda9d0b184c59d1e0a7502a766b880a6d78a73c32270f9fbc49f", "submitted_at": "2026-04-13T10:12:03Z", "updated_at": "2026-04-13T10:12:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.11259v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.11259", "identifiers": {"arxiv": "2604.11259"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.11259"}], "name": "Mobile GUI Agent Privacy Personalization with Trajectory Induced Preference Optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-13", "url": "https://arxiv.org/abs/2604.11259"}
{"aliases": [], "authors": ["Fei Tang", "Zhiqiong Lu", "Boxuan Zhang", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df56b38203bc5ec8d7e871eb2ad9fca9891a049df62b7d72796fe400eb8bce5e", "submitted_at": "2026-04-13T17:52:04Z", "updated_at": "2026-04-13T17:52:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.11784v1"}}, "environments": ["android", "harmonyos", "ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.11784", "identifiers": {"arxiv": "2604.11784"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.11784"}], "name": "ClawGUI: A Unified Framework for Training, Evaluating, and Deploying GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents ClawGUI, a shared harness for GUI-agent reinforcement learning, benchmark evaluation and deployment on mobile devices. It combines parallel virtual and physical environments, process rewards, and hybrid CLI-GUI execution with persistent memory.", "topics": ["reinforcement-learning", "evaluation", "tool-integration", "mobile-navigation"], "updated_at": "2026-04-13", "url": "https://arxiv.org/abs/2604.11784"}
{"aliases": [], "authors": ["Xinyu Jessica Wang", "Haoyue Bai", "Yiyou Sun", "Haorui Wang", "Shuibai Zhang", "Wenjie Hu", "Mya Schroder", "Bilge Mutlu", "Dawn Song", "Robert D Nowak"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0ae917ca6162b9ef0b3757626552725fd54093661e4b0d3cce30ab3eab44732", "submitted_at": "2026-04-13T19:11:42Z", "updated_at": "2026-04-13T19:11:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.11978v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.11978", "identifiers": {"arxiv": "2604.11978"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.11978"}], "name": "The Long-Horizon Task Mirage? Diagnosing Where and Why Agentic Systems Break", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-13", "url": "https://arxiv.org/abs/2604.11978"}
{"aliases": [], "authors": ["Houxing Ren", "Mingjie Zhan", "Zimu Lu", "Ke Wang", "Yunqiao Yang", "Haotian Hou", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "92fa8730caeceac2b61ac0561b06b47c9ec52036a1c37a265acbfb70caff1871", "submitted_at": "2026-04-14T04:47:21Z", "updated_at": "2026-04-14T04:47:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.12282v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.12282", "identifiers": {"arxiv": "2604.12282"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.12282"}], "name": "Towards Robust Real-World Spreadsheet Understanding with Multi-Agent Multi-Format Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2604.12282"}
{"aliases": [], "authors": ["Yulin Chen", "Tri Cao", "Haoran Li", "Yue Liu", "Yibo Li", "Yufei He", "Le Minh Khoi", "Yangqiu Song", "Shuicheng Yan", "Bryan Hooi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64d4323050a5ad0f7265231337eb22a4c0823cee41587099c8bdc5db6b320dea", "submitted_at": "2026-04-14T04:50:35Z", "updated_at": "2026-04-14T04:50:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.12284v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.12284", "identifiers": {"arxiv": "2604.12284"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.12284"}], "name": "WebAgentGuard: A Reasoning-Driven Guard Model for Detecting Prompt Injection Attacks in Web Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-04-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebAgentGuard separates prompt-injection detection from the operating web agent through a parallel multimodal guard. The guard is trained on synthetic visual/textual examples using reasoning-focused supervised fine-tuning followed by reinforcement learning.", "topics": ["safety-security", "data-generation", "imitation-learning", "reinforcement-learning", "evaluation"], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2604.12284"}
{"aliases": [], "authors": ["Guoxin Chen", "Jie Chen", "Lei Chen", "Jiale Zhao", "Fanzhe Meng", "Wayne Xin Zhao", "Ruihua Song", "Cheng Chen", "Ji-Rong Wen", "Kai Jia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5850562978fc6a5b937ad516f3789bfd43fd2e7ead0d8b269eab7a6712f2bc9d", "submitted_at": "2026-04-14T17:55:16Z", "updated_at": "2026-05-26T11:35:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.13018v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.13018", "identifiers": {"arxiv": "2604.13018"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.13018"}], "name": "Toward Autonomous Long-Horizon Engineering for ML Research", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2604.13018"}
{"aliases": [], "authors": ["Himangi Mittal", "Gaurav Mittal", "Nelson Daniel Troncoso", "Yu Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa89e3a09c0ce8b09bdb932f3f00bd2dccf8af4da11a727e675ddf8b2c578146", "submitted_at": "2026-04-14T17:55:46Z", "updated_at": "2026-05-28T06:47:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.13019v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.13019", "identifiers": {"arxiv": "2604.13019"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.13019"}], "name": "PrecisionCUA: Iterative Visual Refinement for Pixel-Precise Cursor Grounding in Code Editors", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies iterative cursor grounding in dense code-editor interfaces, using visual feedback from earlier attempts to correct displacement. PrecisionCUA evaluates this closed-loop process across multiple model families on coding-interface tasks.", "topics": ["grounding", "robustness", "screen-understanding"], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2604.13019"}
{"aliases": [], "authors": ["Zhaoyang Wang", "Qianhui Wu", "Xuchao Zhang", "Chaoyun Zhang", "Wenlin Yao", "Fazle Elahi Faisal", "Baolin Peng", "Si Qin", "Suman Nath", "Qingwei Lin", "Chetan Bansal", "Dongmei Zhang", "Saravan Rajmohan", "Jianfeng Gao", "Huaxiu Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf709a3d7096bc07218206edb1accfdda64036a39ea770e637b3e073dd9a76c8", "submitted_at": "2026-04-14T21:48:15Z", "updated_at": "2026-08-31T07:02:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.13318v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.13318", "identifiers": {"arxiv": "2604.13318"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.13318"}], "name": "WebXSkill: Skill Learning for Autonomous Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebXSkill pairs parameterized action programs with step-level language guidance, mining reusable skills from synthetic browser-agent traces. A URL-based graph supports retrieval, and deployment alternates between direct program execution and guided native planning.", "topics": ["memory", "planning", "web-navigation", "automation"], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2604.13318"}
{"aliases": [], "authors": ["Pengcheng Wang", "Jerry Huang", "Jiarui Yao", "Rui Pan", "Peizhi Niu", "Yaowenqi Liu", "Ruida Wang", "Renhao Lu", "Yuwei Guo", "Tong Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50e3131f22b68895fb83eef3c7c91862182db1843c72ee2e19f03f73af537dad", "submitted_at": "2026-04-14T23:16:25Z", "updated_at": "2026-04-14T23:16:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.13346v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.13346", "identifiers": {"arxiv": "2604.13346"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.13346"}], "name": "AgentSPEX: An Agent SPecification and EXecution Language", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-14", "url": "https://arxiv.org/abs/2604.13346"}
{"aliases": [], "authors": ["Ziwei Wang", "Junjie Zheng", "Leyang Yang", "Sheng Zhou", "Xiaoxuan Tang", "Zhouhua Fang", "Zhiwei Liu", "Dajun Chen", "Yong Li", "Jiajun Bu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54cbfa1e0e7848b6022ae0b52f3a4ca94202fcb47e36e4e70ad8f3c0d36a11d6", "submitted_at": "2026-04-15T05:23:04Z", "updated_at": "2026-04-15T05:23:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.13488v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.13488", "identifiers": {"arxiv": "2604.13488"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.13488"}], "name": "Towards Scalable Lightweight GUI Agents via Multi-role Orchestration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2604.13488"}
{"aliases": [], "authors": ["Renqi Chen", "Zeyin Tao", "Jianming Guo", "Jing Wang", "Zezhou Xu", "Jingzhe Zhu", "Qingqing Sun", "Tianyi Zhang", "Shuai Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "580d3643d55d47160d790bda593ee07279ed6f7fd6ba62b3bfe3491a859db315", "submitted_at": "2026-04-15T06:27:49Z", "updated_at": "2026-04-15T06:27:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.13531v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.13531", "identifiers": {"arxiv": "2604.13531"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.13531"}], "name": "RiskWebWorld: A Realistic Interactive Benchmark for GUI Agents in E-commerce Risk Management", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2604.13531"}
{"aliases": [], "authors": ["Zhengxi Lu", "Fei Tang", "Guangyi Liu", "Kaitao Song", "Xu Tan", "Jin Ma", "Wenqi Zhang", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6e81b448c9c22e2fd25dce519db3fdc0585b0265b94d448751d7a37ed910dfa", "submitted_at": "2026-04-15T12:55:46Z", "updated_at": "2026-04-15T12:55:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.13822v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.13822", "identifiers": {"arxiv": "2604.13822"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.13822"}], "name": "UI-Copilot: Advancing Long-Horizon GUI Automation via Tool-Integrated Policy Optimization", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-Copilot separates GUI execution from on-demand memory retrieval and numerical assistance. Its training procedure treats tool selection and multi-turn task execution as distinct optimization targets while keeping persistent observations outside transient context.", "topics": ["tool-integration", "memory", "planning", "long-horizon", "mobile-navigation"], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2604.13822"}
{"aliases": [], "authors": ["Jiacheng Wang", "Jinchang Hou", "Fabian Wang", "Ping Jian", "Chenfu Bao", "Zhonghou Lv"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "30823c16fe9e10e37d1640ce991d75cd4ddddf3a801b8af401acd885fac76d0b", "submitted_at": "2026-04-15T15:06:01Z", "updated_at": "2026-09-16T02:38:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.13954v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.13954", "identifiers": {"arxiv": "2604.13954"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.13954"}], "name": "HINTBench: Horizon-agent Intrinsic Non-attack Trajectory Benchmark", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-16", "url": "https://arxiv.org/abs/2604.13954"}
{"aliases": [], "authors": ["Subhradip Chakraborty", "Ankur Singh", "Akhilesh R. Jaiswal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e93119d075175dbc389be98eb0d38019687a64b9a362024923458326757e07cc", "submitted_at": "2026-04-15T15:18:33Z", "updated_at": "2026-04-15T15:18:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.13969v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.13969", "identifiers": {"arxiv": "2604.13969"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.13969"}], "name": "GEM3D CIM General Purpose Matrix Computation Using 3D Integrated SRAM eDRAM Hybrid Compute In Memory on Memory Architecture", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2604.13969"}
{"aliases": [], "authors": ["Fei Tang", "Bofan Chen", "Zhengxi Lu", "Tongbo Chen", "Songqin Nong", "Tao Jiang", "Wenhao Xu", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68709f498ebd5d475593ca8ac5bd543a695086d4dffadc113fad908495c234d9", "submitted_at": "2026-04-15T17:32:28Z", "updated_at": "2026-04-15T17:32:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14113v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.14113", "identifiers": {"arxiv": "2604.14113"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14113"}], "name": "UI-Zoomer: Uncertainty-Driven Adaptive Zoom-In for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces UI-Zoomer, a training-free grounding method that uses candidate agreement and token confidence to decide when to crop and retry. Prediction variance determines a separate crop radius for each uncertain interface target.", "topics": ["grounding", "training-free", "efficiency"], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2604.14113"}
{"aliases": [], "authors": ["Yangyue Wang", "Harshvardhan Sikka", "Yash Mathur", "Tony Zhou", "Jinu Nyachhyon", "Pranav Guruprasad"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5dc36ff0c9c4bd39de360824967e53c18fe57562cb8da1462d3e63c0dd348e71", "submitted_at": "2026-04-15T16:39:22Z", "updated_at": "2026-07-01T10:12:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.14262v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.14262", "identifiers": {"arxiv": "2604.14262"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14262"}], "name": "GUI-Perturbed: Domain Randomization Reveals Systematic Brittleness in GUI Grounding Models", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI-Perturbed, which varies screenshot appearance and instruction wording independently to diagnose grounding failures. Experiments on three related model architectures examine relational instructions, browser zoom and augmentation-based fine-tuning.", "topics": ["grounding", "robustness", "evaluation"], "updated_at": "2026-07-01", "url": "https://arxiv.org/abs/2604.14262"}
{"aliases": [], "authors": ["Pan Hao", "Rishi Selvakumaran", "Jacob Sun", "Qianwen Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd3b5e5b9a607023c82150673255eaf2f90c22db86d3c049b2a2fd536aaef17f", "submitted_at": "2026-04-16T06:22:48Z", "updated_at": "2026-04-16T06:22:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14668v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.14668", "identifiers": {"arxiv": "2604.14668"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14668"}], "name": "Beyond Chat and Clicks: GUI Agents for In-Situ Assistance via Live Interface Transformation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "DOMSteer translates help requests and live page context into reversible DOM changes such as control highlighting, contextual tooltips and layout reorganization. The work evaluates in-place GUI assistance with interface tests, use cases and a comparative user study.", "topics": ["human-in-the-loop", "ui-design", "grounding", "evaluation"], "updated_at": "2026-04-16", "url": "https://arxiv.org/abs/2604.14668"}
{"aliases": [], "authors": ["Wenhao Wang", "Haoting Shi", "Mengying Yuan", "Yiquan Lin", "Panrong Tong", "Hanzhang Zhou", "Guangyi Liu", "Pengxiang Zhao", "Yue Wang", "Siheng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1e2b5d7149adcbe634df0bfdc9c34ba436e51e6891df5c4836daea9795061164", "submitted_at": "2026-04-16T12:53:03Z", "updated_at": "2026-04-16T12:53:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14956v1"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.14956", "identifiers": {"arxiv": "2604.14956"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14956"}], "name": "FedGUI: Benchmarking Federated GUI Agents across Heterogeneous Platforms, Devices, and Operating Systems", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces FedGUI, a federated GUI-learning benchmark built from six datasets spanning mobile, web and desktop interfaces. Its experiments separate platform, device, operating-system and data-source heterogeneity when evaluating collaborative training.", "topics": ["evaluation", "robustness"], "updated_at": "2026-04-16", "url": "https://arxiv.org/abs/2604.14956"}
{"aliases": [], "authors": ["Kanzhi Cheng", "Zehao Li", "Zheng Ma", "Nuo Chen", "Jialin Cao", "Qiushi Sun", "Zichen Ding", "Fangzhi Xu", "Hang Yan", "Jiajun Chen", "Anh Tuan Luu", "Jianbing Zhang", "Lewei Lu", "Dahua Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "12c2e5e2ae9bdc871bb63d195194c5387129accf5e8b26d12b21650b92fe82b6", "submitted_at": "2026-04-16T14:53:08Z", "updated_at": "2026-04-16T14:53:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.15093v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.15093", "identifiers": {"arxiv": "2604.15093"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.15093"}], "name": "OpenMobile: Building Open Mobile Agents with Task and Trajectory Synthesis", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents OpenMobile, which builds exploration-based environment memory to synthesize task instructions and switches between learner and expert policies during trajectory collection. The resulting data includes recovery behavior and is tested through mobile-agent training and benchmark-overlap analysis.", "topics": ["data-generation", "imitation-learning", "mobile-navigation"], "updated_at": "2026-04-16", "url": "https://arxiv.org/abs/2604.15093"}
{"aliases": [], "authors": ["Yining Hong", "Yining She", "Eunsuk Kang", "Christopher S. Timperley", "Christian Kästner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f2089729fd2d1dd99e0542d929f8fe0075512e4e230640db291757fa502e5ca", "submitted_at": "2026-04-16T23:18:22Z", "updated_at": "2026-07-05T22:49:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.15579v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.15579", "identifiers": {"arxiv": "2604.15579"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.15579"}], "name": "Don't Make Models Guess Security and Safety: Symbolic Guardrails for Domain-Specific AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-05", "url": "https://arxiv.org/abs/2604.15579"}
{"aliases": [], "authors": ["Jize Wang", "Xuanxuan Liu", "Yining Li", "Songyang Zhang", "Yijun Wang", "Zifei Shan", "Xinyi Le", "Cailian Chen", "Xinping Guan", "Dacheng Tao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "35f0d665f828c4f0f8e4e00bda5d2b83372f190aed8102691e6ae835b6218c44", "submitted_at": "2026-04-17T05:36:00Z", "updated_at": "2026-04-17T05:36:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.15715v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.15715", "identifiers": {"arxiv": "2604.15715"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.15715"}], "name": "GTA-2: Benchmarking General Tool Agents from Atomic Tool-Use to Open-Ended Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-17", "url": "https://arxiv.org/abs/2604.15715"}
{"aliases": [], "authors": ["Haoyue Bai", "Dong Wang", "Long Chen", "Bingguang Hao", "Pengyang Shao", "Yonghui Yang", "Yicheng He", "Chenyi Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3ae241d6d51fe73020d8d39c78ba0fb6ef6ffcb5b9fdb2e43bdb1b6f12931210", "submitted_at": "2026-03-27T02:41:38Z", "updated_at": "2026-03-27T02:41:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.16385v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.16385", "identifiers": {"arxiv": "2604.16385"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.16385"}], "name": "StressWeb: A Diagnostic Benchmark for Web Agent Robustness under Realistic Interaction Variability", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-27", "url": "https://arxiv.org/abs/2604.16385"}
{"aliases": [], "authors": ["Jiachen Qian", "Zhaolu Kang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64e47028e752dd023e71414664a193beba74e1c4f1afe765fdf992d560dbea27", "submitted_at": "2026-04-15T10:46:38Z", "updated_at": "2026-04-15T10:46:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.16515v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2604.16515", "identifiers": {"arxiv": "2604.16515"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.16515"}], "name": "Penny Wise, Pixel Foolish: Bypassing Price Constraints in Multimodal Agents via Visual Adversarial Perturbations", "notes": "Scope is screenshot-based interface understanding and safety. The retained abstract does not establish a specific operating system or live transaction execution; no attack results were independently reproduced.", "published_at": "2026-04-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PriceBlind, a visual perturbation method for controlled screenshot-based evaluation of price-constrained multimodal agents. The reported transfer experiments use a simplified single-turn coordinate-selection protocol.", "topics": ["safety-security", "robustness"], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2604.16515"}
{"aliases": [], "authors": ["Haomin Zhuang", "Hanwen Xing", "Xiangliang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b1788e6fbb4c14a6d993ce329bf7bf05e3a1e6ffbfc7c14a9fa9d36fbf7ef16", "submitted_at": "2026-04-15T18:22:50Z", "updated_at": "2026-04-15T18:22:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.16520v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.16520", "identifiers": {"arxiv": "2604.16520"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.16520"}], "name": "AgentClick: A Skill-Based Human-in-the-Loop Review Layer for Terminal AI Agents", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent supervision interface: the browser UI is for human review of terminal agents; the abstract does not claim that those agents operate graphical interfaces.", "published_at": "2026-04-15", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "AgentClick gives terminal-based agents a browser review interface for plans, drafts, execution traces and editable memory. A local server and skill plugin let users inspect proposed actions and provide structured feedback.", "topics": ["human-in-the-loop", "observability", "workflow", "memory", "tool-integration"], "updated_at": "2026-04-15", "url": "https://arxiv.org/abs/2604.16520"}
{"aliases": [], "authors": ["Zehao Lin", "Xixuan Hao", "Renyu Fu", "Shaobo Cui", "Kai Chen", "Chunyu Li", "Zhiyu Li", "Feiyu Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb75ea30086ecf069d8adda1513a6e906cfcfcedcd4cab01d0986a850aa8009a", "submitted_at": "2026-04-17T06:28:22Z", "updated_at": "2026-09-22T07:33:24Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.16548v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T13:00:08.054642+00:00", "role": "primary", "sha256": "f795535d55a4def0f2ce7dfc41a9d1d22ad38eddf48fb4f1feed98b6b5b66721", "supports": ["summary", "scope", "notes"], "url": "https://arxiv.org/html/2604.16548v3"}], "id": "arxiv:2604.16548", "identifiers": {"arxiv": "2604.16548"}, "kind": "survey", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.16548"}], "name": "A Survey on Long-Term Memory Security in LLM Agents: Attacks, Defenses, and Governance Across the Memory Lifecycle", "notes": "Adjacent memory-security survey. Full-text Sections 3.1 and 3.3 discuss poisoned web observations persisted into later agent actions; the work is broader than GUI agents and does not establish a platform implementation.", "published_at": "2026-04-17", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Organizes security risks and defenses for persistent agent memory into six lifecycle phases and proposes five verifiable governance primitives.", "topics": ["memory", "safety-security", "survey"], "updated_at": "2026-09-22", "url": "https://arxiv.org/abs/2604.16548"}
{"aliases": [], "authors": ["Chao Jin", "Wenkui Yang", "Hao Sun", "Yuqi Liao", "Qianyi Jiang", "Kai Zhou", "Jie Cao", "Ran He", "Huaibo Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b102c9363b1be77273c7fab6f7f5cde74157c646671c93b9503770fe60dd70c", "submitted_at": "2026-04-19T06:55:59Z", "updated_at": "2026-04-19T06:55:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17284v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.17284", "identifiers": {"arxiv": "2604.17284"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.17284"}], "name": "HalluClear: Diagnosing, Evaluating and Mitigating Hallucinations in GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces HalluClear with a GUI hallucination taxonomy, expert-calibrated evaluation workflow and structured-reasoning post-training scheme. It evaluates how these components affect hallucinations, grounding and action fidelity across GUI agents.", "topics": ["robustness", "grounding", "evaluation"], "updated_at": "2026-04-19", "url": "https://arxiv.org/abs/2604.17284"}
{"aliases": [], "authors": ["Hailin Liu", "Eugene Ilyushin", "Jie Ni", "Min Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "279b5bf3affad907645fa2fb34a0017f92ce15a2c7cab5e215013b5f3717097c", "submitted_at": "2026-04-19T18:02:21Z", "updated_at": "2026-04-19T18:02:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17562v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2604.17562", "identifiers": {"arxiv": "2604.17562"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.17562"}], "name": "SafeAgent: A Runtime Protection Architecture for Agentic Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-19", "url": "https://arxiv.org/abs/2604.17562"}
{"aliases": [], "authors": ["Shiquan Zhang", "Tianyi Zhang", "Le Fang", "Simon D'Alfonso", "Hong Jia", "Vassilis Kostakos"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "176f4d68e2f5084bee1938ee1399978052cdce2d2250ef06c16ddbf49c977da1", "submitted_at": "2026-04-20T05:15:14Z", "updated_at": "2026-04-20T05:15:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17817v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "id": "arxiv:2604.17817", "identifiers": {"arxiv": "2604.17817"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.17817"}], "name": "Do LLMs Need to See Everything? A Benchmark and Study of Failures in LLM-driven Smartphone Automation using Screentext vs. Screenshots", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-04-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "DailyDroid compares text-only and text-plus-screenshot agents on everyday Android tasks across multiple apps and difficulty levels. Its failure analysis examines accessibility, observation modality and agent/application behavior.", "topics": ["evaluation", "mobile-navigation", "screen-understanding", "robustness"], "updated_at": "2026-04-20", "url": "https://arxiv.org/abs/2604.17817"}
{"aliases": [], "authors": ["Lingfeng Zhang", "Yongan Sun", "Jinpeng Hu", "Hui Ma", "Yang Ying", "Kuien Liu", "Zenglin Shi", "Meng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "123577fbb4431f22bc18d2d32187744268f9d63b3149e9e1775fb44673771feb", "submitted_at": "2026-04-20T05:19:49Z", "updated_at": "2026-04-21T03:15:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.17821v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.17821", "identifiers": {"arxiv": "2604.17821"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.17821"}], "name": "WebUncertainty: Dual-Level Uncertainty Driven Planning and Reasoning For Autonomous Web Agent", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebUncertainty adapts planning mode to task uncertainty and guides action-level tree search with confidence-derived uncertainty estimates. Its evaluations examine autonomous execution of dynamic, long-horizon web tasks on WebArena and WebVoyager.", "topics": ["planning", "web-navigation", "long-horizon", "robustness"], "updated_at": "2026-04-21", "url": "https://arxiv.org/abs/2604.17821"}
{"aliases": [], "authors": ["Gonzalo Gonzalez-Pumariega", "Saaket Agashe", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a508db9eba8a7db5b52273bff32a16189f5a8eed19d6563e0f84313188a0aad1", "submitted_at": "2026-04-20T05:59:04Z", "updated_at": "2026-04-20T05:59:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17849v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.17849", "identifiers": {"arxiv": "2604.17849"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.17849"}], "name": "On the Reliability of Computer Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-04-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study examines repeated OSWorld executions to separate execution stochasticity, task-specification ambiguity and behavioral variation. Paired task-level tests show why one successful run is insufficient evidence of reliable computer use.", "topics": ["evaluation", "robustness"], "updated_at": "2026-04-20", "url": "https://arxiv.org/abs/2604.17849"}
{"aliases": [], "authors": ["Wentao Shi", "Yu Wang", "Yuyang Zhao", "Yuxin Chen", "Fuli Feng", "Xueyuan Hao", "Xi Su", "Qi Gu", "Hui Su", "Xunliang Cai", "Xiangnan He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "91a9cdb6c8cb124d2e6033f0decc85e0c73bd972984d973ff687bbdc2be21176", "submitted_at": "2026-04-20T13:23:38Z", "updated_at": "2026-04-20T13:23:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.18240v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.18240", "identifiers": {"arxiv": "2604.18240"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.18240"}], "name": "AJ-Bench: Benchmarking Agent-as-a-Judge for Environment-Aware Evaluation", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-04-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "AJ-Bench evaluates judge agents that actively gather evidence from search, data systems and graphical interfaces. Its annotated tasks and trajectories test information acquisition, state verification and process verification.", "topics": ["evaluation", "reward-models", "state-tracking"], "updated_at": "2026-04-20", "url": "https://arxiv.org/abs/2604.18240"}
{"aliases": [], "authors": ["Christy Li", "Sky CH-Wang", "Andi Peng", "Andreea Bobu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9064a2efc0e8952870ac5c498ee597622238c9daa77a3a9b89549eb768c5f240", "submitted_at": "2026-04-20T21:12:40Z", "updated_at": "2026-05-28T03:46:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.18847v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.18847", "identifiers": {"arxiv": "2604.18847"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.18847"}], "name": "Human-Guided Harm Recovery for Computer Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2604.18847"}
{"aliases": [], "authors": ["Wenpeng Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8eb499729268ef287a9b9bf73275f526d47dabf2d7de399beaead2fe54cfd74e", "submitted_at": "2026-04-20T21:36:16Z", "updated_at": "2026-04-20T21:36:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.18860v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.18860", "identifiers": {"arxiv": "2604.18860"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.18860"}], "name": "Temporal UI State Inconsistency in Desktop GUI Agents: Formalizing and Defending Against TOCTOU Attacks on Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-04-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study examines UI changes between a desktop agent’s screenshot observation and action dispatch. Its pre-execution state checks address overlay and focus manipulation, while the reported DOM-injection blind spot limits the defense’s coverage.", "topics": ["safety-security", "robustness", "evaluation", "state-tracking"], "updated_at": "2026-04-20", "url": "https://arxiv.org/abs/2604.18860"}
{"aliases": [], "authors": ["Daniel Shepard", "Robin Salimans"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0291c8077613d0836ee73cfbdca8a5dac82d1c6305e0eca6776a4667bac4b66", "submitted_at": "2026-04-21T00:14:59Z", "updated_at": "2026-04-21T00:14:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.18934v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.18934", "identifiers": {"arxiv": "2604.18934"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.18934"}], "name": "AutomationBench", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-21", "url": "https://arxiv.org/abs/2604.18934"}
{"aliases": [], "authors": ["Xinlin Wang", "Mats Brorsson"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "94f8b4e2524f738da8dbdffa4315f4898d5f664e38c40d6f01f7851582add425", "submitted_at": "2026-04-21T10:05:10Z", "updated_at": "2026-04-21T10:05:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.19299v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "id": "arxiv:2604.19299", "identifiers": {"arxiv": "2604.19299"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.19299"}], "name": "Rethinking Scale: Deployment Trade-offs of Small Language Models under Agent Paradigms", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-21", "url": "https://arxiv.org/abs/2604.19299"}
{"aliases": [], "authors": ["Jincheng Ren", "Siwei Wu", "Yizhi Li", "Kang Zhu", "Shu Xu", "Boyu Feng", "Ruibin Yuan", "Wei Zhang", "Riza Batista-Navarro", "Jian Yang", "Chenghua Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6890d13b68bf0aca6d9fd07839c063526f7583e4c6f1ded10efe316a453e2daf", "submitted_at": "2026-04-21T15:25:54Z", "updated_at": "2026-05-15T13:11:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.19572v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.19572", "identifiers": {"arxiv": "2604.19572"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.19572"}], "name": "A Self-Evolving Framework for Efficient Terminal Agents via Observational Context Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-15", "url": "https://arxiv.org/abs/2604.19572"}
{"aliases": [], "authors": ["Yi Zhong", "Buqiang Xu", "Yijun Wang", "Zifei Shan", "Shuofei Qiao", "Guozhou Zheng", "Ningyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9df86d68ca6178386259235d72db4ec37918cdc06fb80ff44b35282dc1f03c35", "submitted_at": "2026-04-21T16:49:11Z", "updated_at": "2026-05-26T16:14:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.19667v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.19667", "identifiers": {"arxiv": "2604.19667"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.19667"}], "name": "Chat2Workflow: A Benchmark for Generating Executable Visual Workflows with Natural Language", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2604.19667"}
{"aliases": [], "authors": ["Zhiyuan Peng", "Wei Tao", "Xin Yin", "Chenhao Ying", "Yuan Luo", "Yiwen Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "74ec7a508a815781cc7e721aa6c5e351f086f6cab46da48e22c2b58f52af8ce0", "submitted_at": "2026-04-21T17:59:16Z", "updated_at": "2026-09-29T10:11:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.19742v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.19742", "identifiers": {"arxiv": "2604.19742"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.19742"}], "name": "PlayCoder: Making LLM-Generated GUI Code Playable", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-04-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "PlayEval measures whether generated GUI applications remain logically playable, using PlayTester to execute task-oriented GUI playthroughs. PlayCoder closes the loop between repository-aware generation, interactive testing and targeted code repair.", "topics": ["evaluation", "automation", "ui-design", "workflow"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2604.19742"}
{"aliases": [], "authors": ["Jeonghyeon Kim", "Byeongjun Joung", "Junwon Lee", "Joohyung Lee", "Taehoon Min", "Sunjae Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5bb365f901aa4afd09aeb6af45f35bb8b7ee0113d5cec188585062dff0328275", "submitted_at": "2026-04-22T07:27:21Z", "updated_at": "2026-09-07T13:53:30Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.20279v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.20279", "identifiers": {"arxiv": "2604.20279"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.20279"}], "name": "AgentLens: Adaptive Visual Modalities for Human-Agent Interaction in Mobile GUI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-07", "url": "https://arxiv.org/abs/2604.20279"}
{"aliases": [], "authors": ["Xiyang Wu", "Zongxia Li", "Guangyao Shi", "Alexander Duffy", "Tyler Marques", "Matthew Lyle Olson", "Tianyi Zhou", "Dinesh Manocha"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b002b56b12997727d90f7dc6a29aabcf30aeccea90589c2296025bfb6da883b", "submitted_at": "2026-04-22T18:17:17Z", "updated_at": "2026-04-22T18:17:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.20987v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.20987", "identifiers": {"arxiv": "2604.20987"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.20987"}], "name": "Co-Evolving LLM Decision and Skill Bank Agents for Long-Horizon Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-22", "url": "https://arxiv.org/abs/2604.20987"}
{"aliases": [], "authors": ["Wenkai Wang", "Xiyun Li", "Hongcan Guo", "Wenhao Yu", "Tianqing Fang", "Haitao Mi", "Dong Yu", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e15661746ab05f2894e23d28b4675e13919bbcd71407fbc33260d84f330ea5d", "submitted_at": "2026-04-23T04:23:31Z", "updated_at": "2026-04-23T04:23:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.21268v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.21268", "identifiers": {"arxiv": "2604.21268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.21268"}], "name": "Measure Twice, Click Once: Co-evolving Proposer and Visual Critic via Reinforcement Learning for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains a GUI grounding proposer together with a visual critic that evaluates candidate locations rendered on screenshots. An adaptive reinforcement-learning schedule balances their development, using proposal diversity and critic feedback during training.", "topics": ["grounding", "reinforcement-learning", "self-reflection"], "updated_at": "2026-04-23", "url": "https://arxiv.org/abs/2604.21268"}
{"aliases": [], "authors": ["Qijun Han", "Haoqin Tu", "Zijun Wang", "Haoyue Dai", "Yiyang Zhou", "Nancy Lau", "Alvaro A. Cardenas", "Yuhui Xu", "Ran Xu", "Caiming Xiong", "Zeyu Zheng", "Huaxiu Yao", "Yuyin Zhou", "Cihang Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e675476121ab3ef51a8627ccfcc86d76eddde73987ea882398f72e3cb31c5d9", "submitted_at": "2026-04-23T07:42:37Z", "updated_at": "2026-09-24T23:58:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.21375v3"}}, "environments": ["desktop", "linux", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.21375", "identifiers": {"arxiv": "2604.21375"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.21375"}], "name": "VLAA-GUI: Knowing When to Stop, Recover, and Search, A Modular Framework for GUI Automation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "VLAA-GUI combines mandatory checks for visually supported completion, multi-tier loop recovery and on-demand workflow search. Optional coding and grounding agents support execution, with component ablations on Linux and Windows GUI tasks.", "topics": ["robustness", "self-reflection", "tool-integration", "evaluation"], "updated_at": "2026-09-24", "url": "https://arxiv.org/abs/2604.21375"}
{"aliases": [], "authors": ["Bin Wu", "Arastun Mammadli", "Xiaoyu Zhang", "Emine Yilmaz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "705dbf3276834ee43d0ede8c164cec0ee62241581ced106dd3b8d7f0c9339e0c", "submitted_at": "2026-04-24T10:53:54Z", "updated_at": "2026-04-24T10:53:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.22436v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.22436", "identifiers": {"arxiv": "2604.22436"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.22436"}], "name": "AgentSearchBench: A Benchmark for AI Agent Search in the Wild", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-24", "url": "https://arxiv.org/abs/2604.22436"}
{"aliases": [], "authors": ["Fanqing Meng", "Lingxiao Du", "Zijian Wu", "Guanzheng Chen", "Xiangyan Liu", "Jiaqi Liao", "Chonghe Jiang", "Zhenglin Wan", "Jiawei Gu", "Pengfei Zhou", "Rui Huang", "Ziqi Zhao", "Shengyuan Ding", "Ailing Yu", "Bo Peng", "Bowei Xia", "Hao Sun", "Haotian Liang", "Ji Xie", "Jiajun Chen", "Jiajun Song", "Liu Yang", "Ming Xu", "Qionglin Qiu", "Runhao Fu", "Shengfang Zhai", "Shijian Wang", "Tengfei Ma", "Tianyi Wu", "Weiyang Jin", "Yan Wang", "Yang Dai", "Yao Lai", "Youwei Shu", "Yue Liu", "Yunzhuo Hao", "Yuwei Niu", "Jinkai Huang", "Jiayuan Zhuo", "Zhennan Shen", "Linyu Wu", "Hannah Yao", "Charles Chen", "Cihang Xie", "Yuyin Zhou", "Jiaheng Zhang", "Zeyu Zheng", "Mengkang Hu", "Michael Qizhe Shieh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d85c422bb392ec53606435774ec38f359c6423d5eddf3b0b17d110853e28817b", "submitted_at": "2026-04-26T16:05:02Z", "updated_at": "2026-05-05T15:32:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.23781v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.23781", "identifiers": {"arxiv": "2604.23781"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.23781"}], "name": "ClawMark: A Living-World Benchmark for Multi-Turn, Multi-Day, Multimodal Coworker Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-05", "url": "https://arxiv.org/abs/2604.23781"}
{"aliases": [], "authors": ["Hongxin Li", "Yuntao Chen", "Zhaoxiang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3fe6a322737dd58339673bb1c956345ff585d79df6185ce1f9eaa452b37a1e01", "submitted_at": "2026-04-27T01:29:02Z", "updated_at": "2026-04-27T01:29:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.23941v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2604.23941", "identifiers": {"arxiv": "2604.23941"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.23941"}], "name": "GoClick: Lightweight Element Grounding Model for Autonomous GUI Interaction", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2026-04-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies a compact encoder-decoder grounding model and progressive training-data refinement for resource-constrained GUI interaction.", "topics": ["grounding", "efficiency"], "updated_at": "2026-04-27", "url": "https://arxiv.org/abs/2604.23941"}
{"aliases": [], "authors": ["Zheng Wu", "Yi Hua", "Zhaoyuan Huang", "Chenhao Xue", "Yijie Lu", "Pengzhou Cheng", "Zongru Wu", "Lingzhong Dong", "Gongshen Liu", "Xinghao Jiang", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "772420683208561df8493c9ed5d68b341e9e098f5729945b972f3555e976d799", "submitted_at": "2026-04-27T11:44:26Z", "updated_at": "2026-04-27T11:44:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24348v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2604.24348", "identifiers": {"arxiv": "2604.24348"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.24348"}], "name": "OS-SPEAR: A Toolkit for the Safety, Performance,Efficiency, and Robustness Analysis of OS Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-04-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Provides an evaluation toolkit for joint safety, efficiency, and multimodal robustness analysis of operating-system GUI agents.", "topics": ["evaluation", "safety-security", "efficiency", "robustness"], "updated_at": "2026-04-27", "url": "https://arxiv.org/abs/2604.24348"}
{"aliases": [], "authors": ["Hongxin Li", "Xiping Wang", "Jingran Su", "Zheng Ju", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65325912365effe8adbc39301a378d7526c60d1185691f3a89cb8acd4aa53c3f", "submitted_at": "2026-04-27T13:06:27Z", "updated_at": "2026-04-27T13:06:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24441v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2604.24441", "identifiers": {"arxiv": "2604.24441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.24441"}], "name": "AutoGUI-v2: A Comprehensive Multi-Modal GUI Functionality Understanding Benchmark", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "AutoGUI-v2 evaluates interface functionality, grounding and predicted interaction outcomes using hierarchical regions derived through VLM-human annotation. Its tasks span multiple operating systems and expose differences between functional localization, captioning and transition-logic understanding.", "topics": ["screen-understanding", "grounding", "state-tracking", "evaluation", "data-generation"], "updated_at": "2026-04-27", "url": "https://arxiv.org/abs/2604.24441"}
{"aliases": [], "authors": ["Mohammadmehdi Ataei", "Farzaneh Askari", "Kamal Rahimi Malekshan", "Pradeep Kumar Jayaraman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0146569e640895dac03869091fd86b61cb80ac95f118e7bfd80b17a74ae95d55", "submitted_at": "2026-04-27T13:46:41Z", "updated_at": "2026-04-27T13:46:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24479v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.24479", "identifiers": {"arxiv": "2604.24479"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.24479"}], "name": "Zero-to-CAD: Agentic Synthesis of Interpretable CAD Programs at Million-Scale Without Real Data", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-27", "url": "https://arxiv.org/abs/2604.24479"}
{"aliases": [], "authors": ["Lawrence Keunho Jang", "Jing Yu Koh", "Daniel Fried", "Ruslan Salakhutdinov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db70a5736d04d26f10afe3d9a0cfb08955f65f69480a0345ff13b30b2e83df23", "submitted_at": "2026-04-27T20:05:41Z", "updated_at": "2026-04-27T20:05:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24964v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.24964", "identifiers": {"arxiv": "2604.24964"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.24964"}], "name": "Odysseys: Benchmarking Web Agents on Realistic Long Horizon Tasks", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Odysseys, a live-web benchmark of long, multi-site tasks derived from browsing sessions. Graded task rubrics and a rubric-score-per-step measure evaluate both partial progress and trajectory efficiency.", "topics": ["evaluation", "long-horizon", "efficiency"], "updated_at": "2026-04-27", "url": "https://arxiv.org/abs/2604.24964"}
{"aliases": [], "authors": ["Enqi Liu", "Liyuan Pan", "Zhi Gao", "Yan Yang", "Chenrui Shi", "Yang Liu", "Jingrong Wu", "Qing Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54b49f54319922268ba880a1c1e8a27ffa7597e228b18be04caa1a6952a6456a", "submitted_at": "2026-04-28T08:43:58Z", "updated_at": "2026-05-08T04:29:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.25380v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.25380", "identifiers": {"arxiv": "2604.25380"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.25380"}], "name": "Benchmarking and Improving GUI Agents in High-Dynamic Environments", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-04-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces DynamicGUIBench for applications whose interfaces change between actions, together with DynamicUI, an agent that processes screen recordings. The agent selects informative frames, filters inconsistent action reasoning and uses trajectory reflection for subsequent decisions.", "topics": ["video", "state-tracking", "evaluation"], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2604.25380"}
{"aliases": [], "authors": ["Mengyao Du", "Han Fang", "Haokai Ma", "Jiahao Chen", "Kai Xu", "Quanjun Yin", "Ee-Chien Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2610c2658221d7da0bd888bf7b143ca6c4175791058570f6562cfa0c8c3a946", "submitted_at": "2026-04-28T12:32:21Z", "updated_at": "2026-04-28T12:32:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.25562v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.25562", "identifiers": {"arxiv": "2604.25562"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.25562"}], "name": "SnapGuard: Lightweight Prompt Injection Detection for Screenshot-Based Web Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-04-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "SnapGuard detects screenshot-based prompt injection through complementary visual-stability and recovered textual cues. Its evaluation compares detection quality and inference cost across attack and benign-page settings, rather than establishing universal runtime prevention.", "topics": ["safety-security", "screen-understanding", "evaluation", "efficiency"], "updated_at": "2026-04-28", "url": "https://arxiv.org/abs/2604.25562"}
{"aliases": [], "authors": ["Alice Gao", "Weixi Tong", "Rishab Vempati", "Katharina Reinecke", "R. Benjamin Shapiro", "Tianyi Zhang", "Jason Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb6d24f7dc38921da883c4850b37f1c7eb48a03de920356d4ec2e095dbb44b04", "submitted_at": "2026-04-28T18:04:11Z", "updated_at": "2026-04-28T18:04:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.26020v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.26020", "identifiers": {"arxiv": "2604.26020"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.26020"}], "name": "Training Computer Use Agents to Assess the Usability of Graphical User Interfaces", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-04-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "uxCUA learns to assess interface usability by choosing interaction flows, executing them and predicting a numerical usability score. Training uses interactive UIs with usability labels and human preferences, with evaluations on synthetic and real interfaces.", "topics": ["ui-design", "evaluation", "automation"], "updated_at": "2026-04-28", "url": "https://arxiv.org/abs/2604.26020"}
{"aliases": [], "authors": ["Chen Liang", "Xirui Jiang", "Naihao Deng", "Eytan Adar", "Anhong Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5b572e37eefab40a6a21fa4e4589324380ad9f66873a4da5746fc0f46aaeb1f6", "submitted_at": "2026-04-28T22:15:06Z", "updated_at": "2026-08-11T15:48:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.26148v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.26148", "identifiers": {"arxiv": "2604.26148"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.26148"}], "name": "Beyond Screenshots: Evaluating VLMs' Understanding of UI Animations", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-04-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "AniMINT evaluates how VLMs perceive, explain and interpret UI animations using densely annotated videos. The study distinguishes basic motion recognition from higher-level animation meaning and probes motion, context and perceptual cues.", "topics": ["evaluation", "screen-understanding", "video", "ui-design"], "updated_at": "2026-08-11", "url": "https://arxiv.org/abs/2604.26148"}
{"aliases": [], "authors": ["Jinbiao Wei", "Kangqi Ni", "Yilun Zhao", "Guo Gan", "Arman Cohan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68183be37dbd78314aca35f7f1420e5f2f8da8d8c0042a258b45ae03c90606cf", "submitted_at": "2026-04-29T19:59:36Z", "updated_at": "2026-04-29T19:59:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.27151v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.27151", "identifiers": {"arxiv": "2604.27151"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.27151"}], "name": "Step-level Optimization for Efficient Computer-use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes a step-level model cascade that runs a smaller computer-use policy until learned monitors flag stalled progress or semantic drift. Separate history-based stuck detection and milestone checks determine when to escalate for recovery or verification without retraining the larger model.", "topics": ["efficiency", "robustness", "state-tracking", "interaction-history", "long-horizon"], "updated_at": "2026-04-29", "url": "https://arxiv.org/abs/2604.27151"}
{"aliases": [], "authors": ["Fazle Elahi Faisal", "Qianhui Wu", "Baolin Peng", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "276d8a62192c0ad576a1d986bb27d3978963b93466a822858e3313957dd83a81", "submitted_at": "2026-04-29T22:57:35Z", "updated_at": "2026-04-29T22:57:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.27253v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.27253", "identifiers": {"arxiv": "2604.27253"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.27253"}], "name": "AutoSurfer -- Teaching Web Agents through Comprehensive Surfing, Learning, and Modeling", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "AutoSurfer explores website actions breadth-first, retaining page and trajectory knowledge to guide task synthesis and trajectory refinement. It evaluates the resulting training data by fine-tuning a web agent on WebArena and examining synthesized-task diversity.", "topics": ["web-navigation", "data-generation", "imitation-learning", "evaluation"], "updated_at": "2026-04-29", "url": "https://arxiv.org/abs/2604.27253"}
{"aliases": [], "authors": ["Jinchao Li", "Yunxin Li", "Chenrui Zhao", "Zhenran Xu", "Baotian Hu", "Min Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2120860b9c735466a1538ad26c7cda0bc436ab0c6204febdefa08ad66f0036b7", "submitted_at": "2026-04-30T12:13:27Z", "updated_at": "2026-04-30T12:13:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.27776v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.27776", "identifiers": {"arxiv": "2604.27776"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.27776"}], "name": "WindowsWorld: A Process-Centric Benchmark of Autonomous GUI Agents in Professional Cross-Application Environments", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. Desktop use is explicit; no additional operating-system claim is inferred solely from the benchmark name.", "published_at": "2026-04-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WindowsWorld for professional workflows spanning multiple desktop applications, with generated tasks refined through human review. Intermediate subgoals support process-oriented evaluation of cross-application reasoning and execution efficiency.", "topics": ["evaluation", "cross-app", "workflow"], "updated_at": "2026-04-30", "url": "https://arxiv.org/abs/2604.27776"}
{"aliases": [], "authors": ["Junan Hu", "Jian Liu", "Jingxiang Lai", "Jiarui Hu", "Yiwei Sheng", "Shuang Chen", "Jian Li", "Dazhao Du", "Song Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b06f2787f7268d91820270502c5b6d68356a52a551da53c2cdd97fcacb7c1cc9", "submitted_at": "2026-04-30T14:51:49Z", "updated_at": "2026-04-30T14:51:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.27955v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2604.27955", "identifiers": {"arxiv": "2604.27955"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.27955"}], "name": "GUI Agents with Reinforcement Learning: Toward Digital Inhabitants", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-30", "url": "https://arxiv.org/abs/2604.27955"}
{"aliases": [], "authors": ["Jackson Vonderhorst", "Kuangshi Ai", "Haichao Miao", "Shusen Liu", "Chaoli Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e7b07a811df7e74a28f91a5b6c59c69b7ba066bd6c64452a299e678a3187733", "submitted_at": "2026-04-30T15:22:28Z", "updated_at": "2026-06-26T20:31:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.27996v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.27996", "identifiers": {"arxiv": "2604.27996"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.27996"}], "name": "Exploring LLM Agent Designs and Interaction Modalities for Scientific Visualization", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-04-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study compares specialized, computer-use and coding agents on scientific visualization workflows. It examines GUI, CLI, code and tool-call interaction modes, including how persistent memory changes performance across repeated trials.", "topics": ["science", "evaluation", "tool-integration", "memory"], "updated_at": "2026-06-26", "url": "https://arxiv.org/abs/2604.27996"}
{"aliases": [], "authors": ["Ivan Bercovich"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbe66b1ba8ed2314c16c3778377b68eff0ee761bcf1b4ec820d40395b83d6a65", "submitted_at": "2026-04-30T16:37:37Z", "updated_at": "2026-04-30T16:37:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.28093v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.28093", "identifiers": {"arxiv": "2604.28093"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.28093"}], "name": "What Makes a Good Terminal-Agent Benchmark Task: A Guideline for Adversarial, Difficult, and Legible Evaluation Design", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-30", "url": "https://arxiv.org/abs/2604.28093"}
{"aliases": [], "authors": ["Chenxin Li", "Zhengyang Tang", "Mingxin Huang", "Yunlong Lin", "Shijue Huang", "Shengyuan Liu", "Bowen Ye", "Rang Li", "Lei Li", "Benyou Wang", "Yixuan Yuan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43c2d0f942df3a786b2bc96a1668fa8bcbebdac61da7633e3cf2c5370b24570d", "submitted_at": "2026-04-30T17:23:19Z", "updated_at": "2026-05-01T09:39:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.28139v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.28139", "identifiers": {"arxiv": "2604.28139"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.28139"}], "name": "Claw-Eval-Live: A Live Agent Benchmark for Evolving Real-World Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-01", "url": "https://arxiv.org/abs/2604.28139"}
{"aliases": [], "authors": ["Tao Ge", "Baolin Peng", "Hao Cheng", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b4256684e7a9e85ddbfd893c92864853f99532a010a80a340d1a4e168a4412e", "submitted_at": "2026-04-30T17:58:02Z", "updated_at": "2026-04-30T17:58:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.28181v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2604.28181", "identifiers": {"arxiv": "2604.28181"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2604.28181"}], "name": "Synthetic Computers at Scale for Long-Horizon Productivity Simulation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-30", "url": "https://arxiv.org/abs/2604.28181"}
{"aliases": [], "authors": ["Haotian Zhao", "Songlin Zhou", "Yuxin Zhang", "Stephen S. -T. Yau", "Wenyu Zhang", "Lun Tian", "Tianshu Zhu", "Yifeng Huang", "Yucheng Zeng", "Jingnan Gu", "Daxiang Dong", "Jianmin Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45b603508c31d1f520b1acb99f2701521638e09ce3357dcc01ef838cefb07271", "submitted_at": "2026-05-01T05:54:37Z", "updated_at": "2026-05-08T06:22:47Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.00425v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.00425", "identifiers": {"arxiv": "2605.00425"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.00425"}], "name": "AEM: Adaptive Entropy Modulation for Multi-Turn Agentic Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2605.00425"}
{"aliases": [], "authors": ["Shardul Sapkota", "Matthew Jörke", "Zane Sabbagh", "Omar Shaikh", "Grace Wang", "James A. Landay"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "570b8449c666bdfbf90fa380ac52cea7a7351ceafe301ce383ca10e4716600a3", "submitted_at": "2026-05-01T08:13:32Z", "updated_at": "2026-09-09T17:20:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.00497v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.00497", "identifiers": {"arxiv": "2605.00497"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.00497"}], "name": "\"What Are You Really Trying to Do?\": Co-Creating Life Goals from Everyday Computer Use", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-09", "url": "https://arxiv.org/abs/2605.00497"}
{"aliases": [], "authors": ["Michito Takeshita", "Takuro Kawada", "Takumi Ohashi", "Shunsuke Kitada", "Hitoshi Iyatomi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87ef518590e29328ce752791df974c9e2215eac656fcf43fed97dd3e847f3cf7", "submitted_at": "2026-05-01T10:16:00Z", "updated_at": "2026-05-01T10:16:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.00551v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.00551", "identifiers": {"arxiv": "2605.00551"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.00551"}], "name": "A11y-Compressor: A Framework for Enhancing the Efficiency of GUI Agent Observations through Visual Context Reconstruction and Redundancy Reduction", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "A11y-Compressor transforms linearized accessibility trees through modal detection, redundancy reduction and semantic structuring. Its OSWorld evaluation measures both observation-token reduction and task success.", "topics": ["screen-understanding", "efficiency", "grounding", "evaluation"], "updated_at": "2026-05-01", "url": "https://arxiv.org/abs/2605.00551"}
{"aliases": [], "authors": ["Yan Zhang", "Daiqing Wu", "Huawen Shen", "Can Ma", "Yu Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b755ff13cddeaf691dfcd902e5f8a1e73cb0e1899be3ce7c4322ed8be18e654e", "submitted_at": "2026-05-01T13:23:26Z", "updated_at": "2026-05-10T03:40:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.00642v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.00642", "identifiers": {"arxiv": "2605.00642"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.00642"}], "name": "Learn where to Click from Yourself: On-Policy Self-Distillation for GUI Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-10", "url": "https://arxiv.org/abs/2605.00642"}
{"aliases": [], "authors": ["Haowen Hu", "Pengzhou Cheng", "Zheng Wu", "Lingzhong Dong", "Gongshen Liu", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8738f2584f096c1d54555adcf8bdbbb283f0d3178eb1f5e7b7c8a61b2965c15e", "submitted_at": "2026-05-02T02:54:36Z", "updated_at": "2026-05-02T02:54:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.01208v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.01208", "identifiers": {"arxiv": "2605.01208"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.01208"}], "name": "Faithful Mobile GUI Agents with Guided Advantage Estimator", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Faithful-Agent, which first teaches abstention under perturbed evidence and then refines GUI policies with an anchor-based, variance-adaptive advantage estimator. A thought-action consistency reward encourages decisions grounded in displayed evidence and instructions.", "topics": ["reinforcement-learning", "robustness", "mobile-navigation"], "updated_at": "2026-05-02", "url": "https://arxiv.org/abs/2605.01208"}
{"aliases": [], "authors": ["Yuyang Gong", "Zihao Wang", "Jiawei Liu", "XiaoFeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f864419c6793502770467a53b9c6376e43e1a4145b1e9aa8dfe6bd43cae7be3", "submitted_at": "2026-05-02T14:25:26Z", "updated_at": "2026-05-02T14:25:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.01462v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.01462", "identifiers": {"arxiv": "2605.01462"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.01462"}], "name": "LocalAlign: Enabling Generalizable Prompt Injection Defense via Generation of Near-Target Adversarial Examples for Alignment Training", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-02", "url": "https://arxiv.org/abs/2605.01462"}
{"aliases": [], "authors": ["Wenyi Wu", "Sibo Zhu", "Kun Zhou", "Biwei Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c49b46d73d238ca69eca43d72f014356cbf8ef3809d4120888f7ea987962a653", "submitted_at": "2026-05-04T02:58:05Z", "updated_at": "2026-05-04T02:58:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.02168v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2605.02168", "identifiers": {"arxiv": "2605.02168"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.02168"}], "name": "Planner Matters! An Efficient and Unbalanced Multi-agent Collaboration Framework for Long-horizon Planning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-04", "url": "https://arxiv.org/abs/2605.02168"}
{"aliases": [], "authors": ["Ruoqi Liu", "Imran Q. Mohiuddin", "Austin J. Schoeffler", "Kavita Renduchintala", "Ashwin Nayak", "Prasantha L. Vemu", "Shivam C. Vedak", "Kameron C. Black", "John L. Havlik", "Isaac Ogunmola", "Stephen P. Ma", "Roopa Dhatt", "Jonathan H. Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "34a0238a504b60969e4c5112e33888c2bb56b9c390afcf4a223987bea4631318", "submitted_at": "2026-05-04T05:32:25Z", "updated_at": "2026-05-04T05:32:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.02240v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.02240", "identifiers": {"arxiv": "2605.02240"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.02240"}], "name": "PhysicianBench: Evaluating LLM Agents in Real-World EHR Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-04", "url": "https://arxiv.org/abs/2605.02240"}
{"aliases": [], "authors": ["Ruilin Yao", "Shegnwu Xiong", "Tianyu Zou", "Shili Xiong", "Yi Rong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2dc8c2f5a5a97bc372a1bf5ab3fe0537e53a681f808d45877f55cdef6556858c", "submitted_at": "2026-05-04T14:18:46Z", "updated_at": "2026-05-04T14:18:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.02630v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.02630", "identifiers": {"arxiv": "2605.02630"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.02630"}], "name": "AutoFocus: Uncertainty-Aware Active Visual Search for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents AutoFocus, a training-free grounding method that turns sampled coordinates and token uncertainty into spatial region proposals. Shape-aware crops and visual comparison select a target while retaining surrounding interface context.", "topics": ["grounding", "training-free", "screen-understanding"], "updated_at": "2026-05-04", "url": "https://arxiv.org/abs/2605.02630"}
{"aliases": [], "authors": ["Jiateng Liu", "Rushi Wang", "Bingxuan Li", "Kunlun Zhu", "Yifan Shen", "Qingyun Wang", "Ahmed Abbasi", "Denghui Zhang", "Heng Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c910dd10dd4bb33bf211f39ce8c1b928ddd5eb3d1cc875503405af9624a01b72", "submitted_at": "2026-05-04T15:28:59Z", "updated_at": "2026-05-04T15:28:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.02729v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.02729", "identifiers": {"arxiv": "2605.02729"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.02729"}], "name": "Augmenting Interface Usability Heuristics for Reliable Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-04", "url": "https://arxiv.org/abs/2605.02729"}
{"aliases": [], "authors": ["Ruofeng Yang", "Yongcan Li", "Shuai Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d1aecca6b3bc25130604a8e6f0fd55f3ef3c14ca6f0ec645846e50e03575bf97", "submitted_at": "2026-05-04T18:10:15Z", "updated_at": "2026-05-04T18:10:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.03042v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.03042", "identifiers": {"arxiv": "2605.03042"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.03042"}], "name": "ARIS: Autonomous Research via Adversarial Multi-Agent Collaboration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-04", "url": "https://arxiv.org/abs/2605.03042"}
{"aliases": [], "authors": ["Zuoyu Zhang", "Yancheng Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c779d299b31c53eb87285988bc950d9f3e9884d6b61ff6f31946ac21d6422df2", "submitted_at": "2026-05-05T00:21:00Z", "updated_at": "2026-05-05T00:21:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.03242v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.03242", "identifiers": {"arxiv": "2605.03242"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.03242"}], "name": "Enhancing Agent Safety Judgment: Controlled Benchmark Rewriting and Analogical Reasoning for Deceptive Out-of-Distribution Scenarios", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-05", "url": "https://arxiv.org/abs/2605.03242"}
{"aliases": [], "authors": ["Marina Mao", "Alexandr Liu", "Pengbo Li", "Siheng Li", "Bo Zhou", "Xiang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "83b6fdd5c5a6feec2b894a4ad636654989b443c5ab2fa00ef802ae3e9e6004ed", "submitted_at": "2026-05-06T12:02:59Z", "updated_at": "2026-05-06T12:02:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.04811v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.04811", "identifiers": {"arxiv": "2605.04811"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.04811"}], "name": "Tree-based Credit Assignment for Multi-Agent Memory System", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-06", "url": "https://arxiv.org/abs/2605.04811"}
{"aliases": [], "authors": ["Yijun Lu", "Rui Ye", "Yuwen Du", "Jiajun Wang", "Songhua Liu", "Siheng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75e0104df2d00d401e0e3b6af144f18175821455fddd249b9693a2ef847a6718", "submitted_at": "2026-05-06T17:54:16Z", "updated_at": "2026-05-06T17:54:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.05191v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.05191", "identifiers": {"arxiv": "2605.05191"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.05191"}], "name": "LongSeeker: Elastic Context Orchestration for Long-Horizon Search Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-06", "url": "https://arxiv.org/abs/2605.05191"}
{"aliases": [], "authors": ["Sohom Datta", "Alex Nahapetyan", "William Enck", "Alexandros Kapravelos"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ff44befcfdb97b67d1680c97784f9e9665d8d9879c093984f42d06c9516a342", "submitted_at": "2026-05-06T23:19:38Z", "updated_at": "2026-05-06T23:19:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.05509v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2605.05509", "identifiers": {"arxiv": "2605.05509"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.05509"}], "name": "WAAA! Web Adversaries Against Agentic Browsers", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-05-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates agentic browsers against classical web attacks and visual social-engineering lures beyond indirect prompt injection.", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2026-05-06", "url": "https://arxiv.org/abs/2605.05509"}
{"aliases": [], "authors": ["Xiaoming Ren", "Ru Zhen", "Chao Li", "Yang Song", "Qiuxia Hou", "Yanhao Zhang", "Peng Liu", "Qi Qi", "Quanlong Zheng", "Qi Wu", "Zhenyi Liao", "Binqiang Pan", "Haobo Ji", "Haonan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ece93c00c5556dfe50ddd0ba3f52b725f34d1bee18ea3565951bb78343d8ed0", "submitted_at": "2026-05-07T06:58:34Z", "updated_at": "2026-05-21T06:55:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.05765v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:51.952206+00:00", "role": "primary", "sha256": "5cd7fb95d5ee8e8c473e0c9d5b892bf50ee425604c6a320e8e9a12eff17120a1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:51.952206+00:00", "role": "primary", "sha256": "5cd7fb95d5ee8e8c473e0c9d5b892bf50ee425604c6a320e8e9a12eff17120a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:10.101459+00:00", "role": "primary", "sha256": "80375cd810dcda2e9ccca2c6fbee44730de8449972036c548bde312566b20673", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2605.05765v2"}], "id": "arxiv:2605.05765", "identifiers": {"arxiv": "2605.05765"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:51.952206+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.05765"}], "name": "X-OmniClaw Technical Report: A Unified Mobile Agent for Multimodal Understanding and Interaction", "notes": "The report presents demonstrations. Its behavior-cloning component captures runtime skills; it does not establish policy-weight imitation training, and literal trace replay is described as future work. No general benchmark result or release completeness is inferred. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (2 Framework; 3 Omni Perception; 4 Omni Memory; 5.1 Omni Action in App Ecosystem; 5.2 Trajectory Cloned Execution); not an entire-paper reading or reproduction claim.", "published_at": "2026-05-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes an Android agent architecture combining multimodal perception, working and personal memory, hybrid XML/visual grounding, and reusable skills captured from user interactions.", "topics": ["mobile-navigation", "memory", "grounding", "workflow"], "updated_at": "2026-05-21", "url": "https://arxiv.org/abs/2605.05765"}
{"aliases": [], "authors": ["Yiming Lei", "Yuhang Yao", "Yujia Zhang", "Yiqi Wang", "Bo Guan", "Depei Zhu", "Chunhui Wang", "Zhuonan Hao", "Tianyu Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ed187757acdfc599f66f61aa52458c4d6078a4667109da08f84bfeeb793ad65a", "submitted_at": "2026-05-07T07:47:43Z", "updated_at": "2026-07-27T22:01:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.05811v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.05811", "identifiers": {"arxiv": "2605.05811"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.05811"}], "name": "Sheet As Token: A Graph-Enhanced Representation for Multi-Sheet Spreadsheet Understanding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-27", "url": "https://arxiv.org/abs/2605.05811"}
{"aliases": [], "authors": ["Xinglin Wang", "Zishen Liu", "Shaoxiong Feng", "Peiwen Yuan", "Yiwei Li", "Jiayi Shi", "Yueqi Zhang", "Chuyi Tan", "Ji Zhang", "Boyuan Pan", "Yao Hu", "Kan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0d2b76136e6d05a1dc46e2ed305ba24c35038cb40a5c37d917bbc6a693e0314", "submitted_at": "2026-05-07T12:24:21Z", "updated_at": "2026-05-08T14:07:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.06110v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.06110", "identifiers": {"arxiv": "2605.06110"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.06110"}], "name": "On Time, Within Budget: Constraint-Driven Online Resource Allocation for Agentic Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2605.06110"}
{"aliases": [], "authors": ["Di Lu", "Bo Zhang", "Xiyuan Li", "Yongzhi Liao", "Xuewen Dong", "Yulong Shen", "Zhiquan Liu", "Jianfeng Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "427dd3232438f8e51da019cb6834b389e7020099b8af12736dae1aa8c634ff5f", "submitted_at": "2026-05-07T15:08:40Z", "updated_at": "2026-05-07T15:08:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.06393v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.06393", "identifiers": {"arxiv": "2605.06393"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.06393"}], "name": "Constraining Host-Level Abuse in Self-Hosted Computer-Use Agents via TEE-Backed Isolation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-07", "url": "https://arxiv.org/abs/2605.06393"}
{"aliases": [], "authors": ["Borui Zhang", "Bo Zhang", "Bo Wang", "Wenzhao Zheng", "Yuhao Cheng", "Liang Tang", "Yiqiang Yan", "Jie Zhou", "Jiwen Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61c1368f37c2ff80ca67985aff50e2fd26123f795cf2a8c6336a505b3fc02526", "submitted_at": "2026-05-07T17:59:31Z", "updated_at": "2026-05-07T17:59:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.06664v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.06664", "identifiers": {"arxiv": "2605.06664"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.06664"}], "name": "BAMI: Training-Free Bias Mitigation in GUI Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-07", "url": "https://arxiv.org/abs/2605.06664"}
{"aliases": [], "authors": ["Nuno Batista", "Ana Morgado", "Oscar Ferraz", "Sagar Silva Pratapsi", "Jorge Lobo", "Gabriel Falcao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b17a28ceff743f76bb6d63c1d2f672c04918ce0875afd12c9cb1391029034b87", "submitted_at": "2026-05-07T11:26:09Z", "updated_at": "2026-10-14T15:02:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.06727v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.06727", "identifiers": {"arxiv": "2605.06727"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.06727"}], "name": "Medical Imaging Classification with Cold-Atom Reservoir Computing using Auto-Encoders and Surrogate-Driven Training", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2605.06727"}
{"aliases": [], "authors": ["Oğuzhan Fatih Kar", "Roman Bachmann", "Yuanzheng Gong", "Anders Boesen Lindbo Larsen", "Afshin Dehghan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2e1fb158dad88d32f8754ac27f12346ba8fbe5f9a5e736fba3e81390071895a3", "submitted_at": "2026-05-07T17:17:10Z", "updated_at": "2026-05-07T17:17:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.06761v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.06761", "identifiers": {"arxiv": "2605.06761"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.06761"}], "name": "Weblica: Scalable and Reproducible Training Environments for Visual Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Weblica, which combines HTTP-level capture and replay with website-grounded environment synthesis. The resulting interactive web environments support reinforcement learning and tests of navigation transfer and inference-time scaling.", "topics": ["sandbox", "data-generation", "reinforcement-learning"], "updated_at": "2026-05-07", "url": "https://arxiv.org/abs/2605.06761"}
{"aliases": [], "authors": ["Yubin Kim", "Chanwoo Park", "Taehan Kim", "Eugene Park", "Samuel Schmidgall", "Salman Rahman", "Chunjong Park", "Cynthia Breazeal", "Xin Liu", "Hamid Palangi", "Hae Won Park", "Daniel McDuff"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0855d6ac9734db06790a6a1b5ffa4bce92c1e3abb353c33e9e3d613134b87a8", "submitted_at": "2026-05-08T00:48:45Z", "updated_at": "2026-05-08T00:48:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.07073v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.07073", "identifiers": {"arxiv": "2605.07073"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.07073"}], "name": "TeamBench: Evaluating Agent Coordination under Enforced Role Separation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2605.07073"}
{"aliases": [], "authors": ["Zejian Chen", "Zhanyuan Liu", "Chaozhuo Li", "Mengxiang Han", "Songyang Liu", "Litian Zhang", "Feng Gao", "Yiming Hei", "Xi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8d7751a82a92b1864f1afa63a2f2019c6606875767abd1f59815da55b1cd8a32", "submitted_at": "2026-05-08T01:38:46Z", "updated_at": "2026-05-08T01:38:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.07110v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.07110", "identifiers": {"arxiv": "2605.07110"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.07110"}], "name": "Securing Computer-Use Agents: A Unified Architecture-Lifecycle Framework for Deployment-Grounded Reliability", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2605.07110"}
{"aliases": [], "authors": ["Donguk Kwon", "Dongha Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6bf74d53b22530951788bd329167f5ba9474827c712590ae9c6d316fac4d62bc", "submitted_at": "2026-05-08T02:11:23Z", "updated_at": "2026-05-08T02:11:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.07134v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.07134", "identifiers": {"arxiv": "2605.07134"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.07134"}], "name": "Region4Web: Rethinking Observation Space Granularity for Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Reorganizes webpage accessibility trees into functional regions through hierarchical decomposition and semantic abstraction. PageDigest supplies a persistent region-level observation to web agents, evaluated on WebArena.", "topics": ["screen-understanding", "web-navigation", "efficiency"], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2605.07134"}
{"aliases": [], "authors": ["Yingli Zhou", "Wang Shu", "Yaodong Su", "Wenchuan Du", "Yixiang Fang", "Xuemin Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "471b7b3dc9c54858b43a7672f42938e20d0008ecc0c5a32d7412e0907444c1aa", "submitted_at": "2026-05-08T07:10:26Z", "updated_at": "2026-05-26T05:21:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.07358v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.07358", "identifiers": {"arxiv": "2605.07358"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.07358"}], "name": "A Comprehensive Survey on Agent Skills: Taxonomy, Techniques, and Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2605.07358"}
{"aliases": [], "authors": ["Yubin Wu", "Zicheng Cai", "Liping Ning", "Hua Wang", "Zhi Chen", "Yaohua Tang", "Hao Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea5186a99b7dbc7860396835c8adf638c937dbcc673b955d51d85bc5fe7dbb9a", "submitted_at": "2026-05-08T09:38:29Z", "updated_at": "2026-10-08T08:52:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.07505v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.07505", "identifiers": {"arxiv": "2605.07505"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.07505"}], "name": "LiteGUI: Lightweight GUI Agents via Multi-Solution Guided Distillation and Dual-Level Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2605.07505"}
{"aliases": [], "authors": ["Pierluca D'Oro", "Sneha Silwal", "William Wong", "Yuxuan Sun", "Fanyi Xiao", "Manchen Wang", "Eric Gan", "Allen Bolourchi", "Joseph Tighe"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "803dd9a5acac5754110a770a21ac7d882339ea0e16a9852a84a657da20355a25", "submitted_at": "2026-05-07T22:43:43Z", "updated_at": "2026-05-07T22:43:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.08261v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.08261", "identifiers": {"arxiv": "2605.08261"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.08261"}], "name": "Computer Use at the Edge of the Statistical Precipice", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-07", "url": "https://arxiv.org/abs/2605.08261"}
{"aliases": [], "authors": ["Zhichao Liu", "Wenbo Pan", "Haining Yu", "Ge Gao", "Tianqing Zhu", "Xiaohua Jia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "204e3c2627288c7ecd731c7b3c473c34cf9eec976df0a4c8e38b66ed2980f206", "submitted_at": "2026-05-08T14:06:03Z", "updated_at": "2026-05-08T14:06:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.08310v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.08310", "identifiers": {"arxiv": "2605.08310"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.08310"}], "name": "WebTrap: Stealthy Mid-Task Hijacking of Browser Agents During Navigation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2605.08310"}
{"aliases": [], "authors": ["Junwei Liao", "Haoting Shi", "Ruiwen Zhou", "Jiaqian Wang", "Shengtao Zhang", "Wei Zhang", "Ying Wen", "Zhiyu Li", "Feiyu Xiong", "Bo Tang", "Weinan Zhang", "Muning Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b67992ce11d2d99c9bfc7d1ce121da948c35b5811fec0d94a728e6a937478a37", "submitted_at": "2026-05-08T18:30:24Z", "updated_at": "2026-10-12T11:08:11Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2605.08374v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.08374", "identifiers": {"arxiv": "2605.08374"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.08374"}], "name": "MemQ: Integrating Q-Learning into Self-Evolving Memory Agents over Provenance DAGs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2605.08374"}
{"aliases": [], "authors": ["Xinyu Li", "Ronghui Mu", "Lin Li", "Tianjin Huang", "Gaojie Jin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f2e96c13c1224487c845295f42038de867f0594b93f80c4184777a16edd824d6", "submitted_at": "2026-05-09T10:55:11Z", "updated_at": "2026-08-02T00:52:12Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.08876v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2605.08876", "identifiers": {"arxiv": "2605.08876"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.08876"}], "name": "OTora: A Unified Red Teaming Framework for Reasoning-Level Denial-of-Service in LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-02", "url": "https://arxiv.org/abs/2605.08876"}
{"aliases": [], "authors": ["Kechen Fang", "Yihua Qin", "Chongyi Wang", "Wenshuo Ma", "Tianyu Yu", "Yuan Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "872edcb58b93563c394673db5bc09e2acd19d28061fa239a0e1bba1b2af6d7bb", "submitted_at": "2026-05-09T15:10:26Z", "updated_at": "2026-05-09T15:10:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.08985v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.08985", "identifiers": {"arxiv": "2605.08985"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.08985"}], "name": "LLaVA-UHD v4: What Makes Efficient Visual Encoding in MLLMs?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-09", "url": "https://arxiv.org/abs/2605.08985"}
{"aliases": [], "authors": ["Guibin Zhang", "Yaohui Ling", "Fanci Meng", "Kun Wang", "Shuicheng Yan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "729ff6b37da62bb8a9a911865554520e3b29219d9735e087c3186705428d7b05", "submitted_at": "2026-05-10T04:31:23Z", "updated_at": "2026-05-10T04:31:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.09317v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.09317", "identifiers": {"arxiv": "2605.09317"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.09317"}], "name": "Mem-W: Latent Memory-Native GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Mem-W, which compresses previous trajectories and current-session segments into latent memory tokens used directly by the GUI policy. Self-distillation and outcome supervision train this memory representation for web and mobile navigation.", "topics": ["memory", "long-horizon", "imitation-learning"], "updated_at": "2026-05-10", "url": "https://arxiv.org/abs/2605.09317"}
{"aliases": [], "authors": ["Yilin Zhang", "Yingkai Hua", "Chunyu Wei", "Xin Wang", "Yueguo Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b57d784c70b50d2c25055c6d139c0a5acfc97da3a4effbd63536b72d6df38a69", "submitted_at": "2026-05-10T12:11:52Z", "updated_at": "2026-05-10T12:11:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.09497v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.09497", "identifiers": {"arxiv": "2605.09497"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.09497"}], "name": "Don't Click That: Teaching Web Agents to Resist Deceptive Interfaces", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "DUDE combines asymmetric hybrid-reward learning with summaries of deceptive-interface failures to guide safer web actions. Its RUC benchmark evaluates deception susceptibility alongside ordinary task performance across interface scenarios.", "topics": ["safety-security", "reinforcement-learning", "evaluation", "robustness"], "updated_at": "2026-05-10", "url": "https://arxiv.org/abs/2605.09497"}
{"aliases": [], "authors": ["Weikai Xu", "Kun Huang", "Yunren Feng", "Jiaxing Li", "Yuhan Chen", "Yuxuan Liu", "Zhizheng Jiang", "Heng Qu", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Xiaolin Hu", "Bo An"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4374ec1b7c44804bb78ea56bd0aa3214083419f37780d79c5fcfb68468727bbf", "submitted_at": "2026-05-11T10:49:31Z", "updated_at": "2026-05-22T05:43:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.10347v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.10347", "identifiers": {"arxiv": "2605.10347"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.10347"}], "name": "How Mobile World Model Guides GUI Agents?", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares mobile world models that predict state changes as delta text, full text, images or renderable code. Downstream Android benchmarks examine synthetic training trajectories and test-time guidance, including limits of post-action reflection.", "topics": ["world-models", "data-generation", "evaluation"], "updated_at": "2026-05-22", "url": "https://arxiv.org/abs/2605.10347"}
{"aliases": [], "authors": ["Shuangrui Ding", "Xuanlang Dai", "Long Xing", "Shengyuan Ding", "Ziyu Liu", "Yang JingYi", "Penghui Yang", "Zhixiong Zhang", "Xilin Wei", "Xinyu Fang", "Yubo Ma", "Haodong Duan", "Jing Shao", "Jiaqi Wang", "Dahua Lin", "Kai Chen", "Yuhang Zang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b82c3cce7a2b264b2bc1896a4ebe21eaedd5b343c4419c5aafa823f689073ab", "submitted_at": "2026-05-11T17:49:43Z", "updated_at": "2026-05-11T17:49:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.10912v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.10912", "identifiers": {"arxiv": "2605.10912"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.10912"}], "name": "WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-11", "url": "https://arxiv.org/abs/2605.10912"}
{"aliases": [], "authors": ["Chiyeong Heo", "Jaechang Kim", "Junhyuk Kwon", "Hoyoung Kim", "Dongmin Park", "Jonghyun Lee", "Jungseul Ok"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0bd75e6bf9d5a8185a2204cdd99d42f9d96c102f7297abcadb9f8425e9246349", "submitted_at": "2026-05-08T10:57:19Z", "updated_at": "2026-05-08T10:57:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.10966v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.10966", "identifiers": {"arxiv": "2605.10966"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.10966"}], "name": "MMTB: Evaluating Terminal Agents on Multimedia-File Tasks", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent computer-automation benchmark using terminal workflows, explicitly separated from graphical-interface control.", "published_at": "2026-05-08", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "MultiMedia-TerminalBench evaluates agents that use terminal tools to work with audio and video files. The Terminus-MM harness adds multimedia perception to study how accessible evidence changes executable workflows and task outcomes.", "topics": ["evaluation", "automation", "video", "tool-integration"], "updated_at": "2026-05-08", "url": "https://arxiv.org/abs/2605.10966"}
{"aliases": [], "authors": ["Amirhossein Abaskohi", "Yuhang He", "Peter West", "Giuseppe Carenini", "Pranit Chawla", "Vibhav Vineet"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "422d2fdf2d7f665df1a5672e03d837ba41fe9c6832d64c40317f77481f642130", "submitted_at": "2026-05-11T20:27:54Z", "updated_at": "2026-06-05T22:18:46Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.11212v3"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.11212", "identifiers": {"arxiv": "2605.11212"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.11212"}], "name": "ReVision: Scaling Computer-Use Agents via Temporal Visual Redundancy Reduction", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains GUI agents with a learned selector that removes repeated visual patches across consecutive screenshots while preserving spatial structure. ReVision evaluates this history representation on OSWorld, WebTailBench and AgentNetBench.", "topics": ["efficiency", "interaction-history", "screen-understanding"], "updated_at": "2026-06-05", "url": "https://arxiv.org/abs/2605.11212"}
{"aliases": [], "authors": ["Wei Wu", "Ziyang Xu", "Zeyu Zhang", "Yang Zhao", "Hao Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9dc03b7b871de159bd00756b6ce36866bdbab8acba35bc5068b4687f7f975874", "submitted_at": "2026-05-12T00:32:24Z", "updated_at": "2026-05-12T00:32:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.11363v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.11363", "identifiers": {"arxiv": "2605.11363"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.11363"}], "name": "PresentAgent-2: Towards Generalist Multimodal Presentation Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-12", "url": "https://arxiv.org/abs/2605.11363"}
{"aliases": [], "authors": ["Jiarui Jin", "Zexuan Yan", "Shijian Wang", "Wenxiang Jiao", "Yuan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b3fe14d99cff94053606e8bad185dac5f57e098d7c57463b7bef301d1ed420f8", "submitted_at": "2026-05-12T08:14:15Z", "updated_at": "2026-06-04T05:47:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.11732v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.11732", "identifiers": {"arxiv": "2605.11732"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.11732"}], "name": "AgentDisCo: Towards Disentanglement and Collaboration in Open-ended Deep Research Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2605.11732"}
{"aliases": [], "authors": ["Yixu Feng", "Zinan Zhao", "Yanxiang Ma", "Chenghao Xia", "Chengbin Du", "Yunke Wang", "Chang Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bba202e1186bf4fa6e08efbc4fbffa3e56391e3d3299b54670361e612ea9f21a", "submitted_at": "2026-05-12T09:08:42Z", "updated_at": "2026-05-17T20:26:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.11817v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "id": "arxiv:2605.11817", "identifiers": {"arxiv": "2605.11817"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.11817"}], "name": "See What Matters: Differentiable Grid Sample Pruning for Generalizable Vision-Language-Action Model", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-17", "url": "https://arxiv.org/abs/2605.11817"}
{"aliases": [], "authors": ["Sina Mavali", "David Pape", "Jonathan Evertz", "Samira Abedini", "Devansh Srivastav", "Thorsten Eisenhofer", "Sahar Abdelnabi", "Lea Schönherr"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "66c89136962543cad391fca4a62fa0e69ee234e80d8b0e9b69a2d86fff63ca5d", "submitted_at": "2026-05-12T15:06:15Z", "updated_at": "2026-05-12T15:06:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12233v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.12233", "identifiers": {"arxiv": "2605.12233"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.12233"}], "name": "No More, No Less: Task Alignment in Terminal Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-12", "url": "https://arxiv.org/abs/2605.12233"}
{"aliases": [], "authors": ["Zerui Qin", "Sheng Yue", "Xingyuan Hua", "Yongjian Fu", "Ju Ren"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "388eac2721929fc8211f40cc4edde61be0b6e6145fa60fbaa444ec80822fcc36", "submitted_at": "2026-05-12T15:48:44Z", "updated_at": "2026-05-12T15:48:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12294v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.12294", "identifiers": {"arxiv": "2605.12294"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12294"}], "name": "Executable Agentic Memory for GUI Agent", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Executable Agentic Memory stores GUI routines in a knowledge graph built through state-aware exploration and action-group mining. A learned value model guides tree search over this executable memory for retrieval-based planning, with AndroidWorld and efficiency evaluation.", "topics": ["memory", "planning", "long-horizon", "efficiency"], "updated_at": "2026-05-12", "url": "https://arxiv.org/abs/2605.12294"}
{"aliases": [], "authors": ["Xuhao Hu", "Xi Zhang", "Haiyang Xu", "Kyle Qiao", "Jingyi Yang", "Xuanjing Huang", "Jing Shao", "Ming Yan", "Jieping Ye"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "09128e90b925ac0cd2cf6e967853180bbef1e97a0960c22065f1b0b50c30d58b", "submitted_at": "2026-05-12T17:57:04Z", "updated_at": "2026-09-19T14:53:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.12481v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.12481", "identifiers": {"arxiv": "2605.12481"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12481"}], "name": "ToolCUA: Towards Optimal GUI-Tool Path Orchestration for Computer Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-05-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "ToolCUA synthesizes interleaved GUI/tool trajectories and trains agents to choose between graphical actions and higher-level tools. Staged supervised and reinforcement learning culminate in online hybrid-environment training with a reward for efficient execution paths.", "topics": ["tool-integration", "data-generation", "imitation-learning", "reinforcement-learning", "planning", "efficiency"], "updated_at": "2026-09-19", "url": "https://arxiv.org/abs/2605.12481"}
{"aliases": [], "authors": ["Di Wu", "Zixiang Ji", "Asmi Kawatkar", "Bryan Kwan", "Jia-Chen Gu", "Nanyun Peng", "Kai-Wei Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5893d1d9205776667a1e1d89720ec3baf55fafe80cc9c6201ca97088f44a10ac", "submitted_at": "2026-05-12T17:59:34Z", "updated_at": "2026-05-12T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12493v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.12493", "identifiers": {"arxiv": "2605.12493"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12493"}], "name": "LongMemEval-V2: Evaluating Long-Term Agent Memory Toward Experienced Colleagues", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-05-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "LongMemEval-V2 asks memory systems to recover environment knowledge from long web-agent histories for question answering. Its AgentRunbook variants contrast indexed knowledge pools with coding-agent retrieval from stored trajectory files.", "topics": ["memory", "interaction-history", "state-tracking", "workflow", "evaluation"], "updated_at": "2026-05-12", "url": "https://arxiv.org/abs/2605.12493"}
{"aliases": [], "authors": ["Miaosen Zhang", "Xiaohan Zhao", "Zhihong Tan", "Zhou Huoshen", "Yijia Fan", "Yifan Yang", "Kai Qiu", "Bei Liu", "Justin Wagle", "Chenzhong Yin", "Mingxi Cheng", "Ji Li", "Qi Dai", "Chong Luo", "Xu Yang", "Xin Geng", "Baining Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1fb57ba3cd858513d1da58d64a7c85388f12e331c9ea9f89204be5f43e6fec38", "submitted_at": "2026-05-12T17:59:58Z", "updated_at": "2026-05-12T17:59:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12501v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.12501", "identifiers": {"arxiv": "2605.12501"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12501"}], "name": "Covering Human Action Space for Computer Use: Data Synthesis and Benchmark", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CUActSpot to evaluate clicks, drags, drawing and other screen interactions across GUI, text, tables, canvases and natural images. A renderer-based pipeline records geometry and screenshots, then generates matching instructions and action traces for training a grounding model.", "topics": ["grounding", "data-generation", "evaluation", "action-prediction"], "updated_at": "2026-05-12", "url": "https://arxiv.org/abs/2605.12501"}
{"aliases": [], "authors": ["Jiaping Lin", "Fei Shen", "Junzhe Li", "Ping Nie", "Fei Yu", "Ming Li", "Haizhou Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cb60b76fca3a8825030b1a31a7b97ff31e06f30ac3852f243efe440f965e5578", "submitted_at": "2026-05-10T07:04:07Z", "updated_at": "2026-05-10T07:04:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12549v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.12549", "identifiers": {"arxiv": "2605.12549"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12549"}], "name": "What Happens Before Decoding? Prefill Determines GUI Grounding in VLMs", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-05-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Re-Prefill adds an attention-guided second prefill pass that revisits candidate screen regions before coordinate generation. The study evaluates this training-free intervention across GUI-grounding models and benchmarks.", "topics": ["grounding", "training-free", "screen-understanding", "evaluation"], "updated_at": "2026-05-10", "url": "https://arxiv.org/abs/2605.12549"}
{"aliases": [], "authors": ["Dylan Zhang", "Yanshan Lin", "Zhengkun Wu", "Yihang Sun", "Bingxuan Li", "Dianqi Li", "Hao Peng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bcaffe303a8eeecbd60973ec8ed635f8a7bbe0eca7c1024d5de1063265010f69", "submitted_at": "2026-05-13T04:15:50Z", "updated_at": "2026-08-29T19:20:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.12978v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.12978", "identifiers": {"arxiv": "2605.12978"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.12978"}], "name": "Useful Memories Become Faulty When Continuously Updated by LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-29", "url": "https://arxiv.org/abs/2605.12978"}
{"aliases": [], "authors": ["Kebin Contreras", "Carlos Hinojosa", "Jorge Bacca", "Bernard Ghanem"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f7dcca695e0d360449d87220082826faffefcd9b0e543b60718f00e60db40739", "submitted_at": "2026-05-13T14:58:24Z", "updated_at": "2026-06-10T08:07:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.13631v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.13631", "identifiers": {"arxiv": "2605.13631"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.13631"}], "name": "ProjGuard: Safety Monitoring for Computer-Use Agents via Low-Dimensional Projections", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-10", "url": "https://arxiv.org/abs/2605.13631"}
{"aliases": [], "authors": ["Yuxiang Lai", "Peng Xia", "Haonian Ji", "Kaiwen Xiong", "Kaide Zeng", "Jiaqi Liu", "Fang Wu", "Jike Zhong", "Zeyu Zheng", "Cihang Xie", "Huaxiu Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "35a45f26bde991b6b2dee8248bf514939aa133577dc9efcfa5d6ae897229d581", "submitted_at": "2026-05-13T21:34:08Z", "updated_at": "2026-05-18T05:36:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.14133v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.14133", "identifiers": {"arxiv": "2605.14133"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.14133"}], "name": "ClawForge: Generating Executable Interactive Benchmarks for Command-Line Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-18", "url": "https://arxiv.org/abs/2605.14133"}
{"aliases": [], "authors": ["Julien Piet", "Annabella Chow", "Yiwei Hou", "Muxi Lyu", "Sylvie Venuto", "Jinhao Zhu", "Raluca Ada Popa", "David Wagner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "572be49190554bbfb6accac4c87ef91608f88e90d06250f229db2922e19fcd56", "submitted_at": "2026-05-14T02:48:57Z", "updated_at": "2026-05-14T02:48:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.14290v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.14290", "identifiers": {"arxiv": "2605.14290"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.14290"}], "name": "Web Agents Should Adopt the Plan-Then-Execute Paradigm", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced. Directly analyzes browser-agent prompt-injection boundaries and WebArena task structure, but advocates typed website operations rather than demonstrating a new GUI-control policy.", "published_at": "2026-05-14", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Argues for compiling web tasks into predefined execution graphs before exposing an agent to untrusted runtime content. Its WebArena task analysis identifies typed, semantically complete website tools as a prerequisite and distinguishes this interface proposal from reactive click-level control.", "topics": ["planning", "safety-security", "tool-integration"], "updated_at": "2026-05-14", "url": "https://arxiv.org/abs/2605.14290"}
{"aliases": [], "authors": ["Isaac Lera", "Carlos Guerrero"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5312785e1b1ba5b3b7566b5536a155e741b360438ace1186a567aa2cd794e3c0", "submitted_at": "2026-05-14T10:06:49Z", "updated_at": "2026-05-14T10:06:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.14649v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.14649", "identifiers": {"arxiv": "2605.14649"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.14649"}], "name": "Multi-objective application placement in fog computing using graph neural network-based reinforcement learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-14", "url": "https://arxiv.org/abs/2605.14649"}
{"aliases": [], "authors": ["Weimin Xiong", "Shuhao Gu", "Bowen Ye", "Zihao Yue", "Lei Li", "Feifan Song", "Sujian Li", "Hao Tian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0403b56b03da2f8264cbfbd6083c37bfdd32e47fdd2e39b4543d9ef2070c610", "submitted_at": "2026-05-14T12:14:24Z", "updated_at": "2026-05-14T12:14:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.14747v1"}, "publication": {"authors": ["Weimin Xiong", "Shuhao Gu", "Bowen Ye", "Zihao Yue", "Lei Li", "Feifan Song", "Sujian Li", "Hao Tian"], "date_basis": "Explicit publisher metadata: 2026/09/29; precision=day.", "date_parts": [2026, 9, 29], "doi": null, "event_dates": "06--11 Jul 2026", "evidence_urls": ["https://proceedings.mlr.press/v306/xiong26c.html"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "This work already exists as metadata_verified arxiv:2605.14747. The official publication abstract still promises a future release; current release status is not inferred from publication. The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "140253-140283", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v306/main/assets/xiong26c/xiong26c.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "Video2GUI: Synthesizing Large-Scale Interaction Trajectories for Generalized GUI Agent Pretraining", "type": "conference-paper", "url": "https://proceedings.mlr.press/v306/xiong26c.html", "venue": "Proceedings of the 43rd International Conference on Machine Learning", "volume": "306"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "fea22076271c2c45814795b5b0c04b6f6a2246803d82785af20496c4787a5c2e", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v306/xiong26c.html"}], "id": "arxiv:2605.14747", "identifiers": {"arxiv": "2605.14747"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.14747"}, {"rel": "publication", "url": "https://proceedings.mlr.press/v306/xiong26c.html"}, {"rel": "publication_pdf", "url": "https://raw.githubusercontent.com/mlresearch/v306/main/assets/xiong26c/xiong26c.pdf"}], "name": "Video2GUI: Synthesizing Large-Scale Interaction Trajectories for Generalized GUI Agent Pretraining", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-14", "url": "https://arxiv.org/abs/2605.14747"}
{"aliases": [], "authors": ["William Lugoloobi", "Samuelle Marro", "Jabez Magomere", "Joss Wright", "Chris Russell"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "69afd6b998ab3d284ef83513914b3892c6dc8d2a383f332467176b4dd7b03c7d", "submitted_at": "2026-05-14T12:55:19Z", "updated_at": "2026-05-14T12:55:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.14786v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.14786", "identifiers": {"arxiv": "2605.14786"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.14786"}], "name": "Known By Their Actions: Fingerprinting LLM Browser Agents via UI Traces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-14", "url": "https://arxiv.org/abs/2605.14786"}
{"aliases": [], "authors": ["Tri Cao", "Yulin Chen", "Hieu Cao", "Yibo Li", "Khoi Le", "Thong Nguyen", "Yuexin Li", "Yufei He", "Yue Liu", "Shuicheng Yan", "Bryan Hooi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "618a914bb7605cc542b748e682166d7456221bccc38fcbe43c66ced6c3321832", "submitted_at": "2026-05-14T16:26:27Z", "updated_at": "2026-05-14T16:26:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.15030v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.15030", "identifiers": {"arxiv": "2605.15030"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.15030"}], "name": "WARD: Adversarially Robust Defense of Web Agents Against Prompt Injections", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-14", "url": "https://arxiv.org/abs/2605.15030"}
{"aliases": [], "authors": ["Baolin Peng", "Wenlin Yao", "Qianhui Wu", "Hao Cheng", "Xiao Yu", "Rui Yang", "Tao Ge", "Alessandro Sordoni", "Xingdi Yuan", "Yelong Shen", "Pengcheng He", "Tong Zhang", "Zhou Yu", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d22cca677c0d8325233693dde0c178372c3fe0840483721702b394817815f6c7", "submitted_at": "2026-05-14T16:35:12Z", "updated_at": "2026-07-30T16:17:23Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.15040v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2605.15040", "identifiers": {"arxiv": "2605.15040"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.15040"}], "name": "Orchard: An Open-Source Agentic Modeling Framework", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-30", "url": "https://arxiv.org/abs/2605.15040"}
{"aliases": [], "authors": ["Yichao Liu", "Huawen Shen", "Liu Yu", "Shiyu Liu", "Zeyu Chen", "Yu Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd85c0d0a8b6bc5c687bc301b1b84d6a019c5713012890e0a23c08ae2d83a1e6", "submitted_at": "2026-05-15T02:27:41Z", "updated_at": "2026-05-15T02:27:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.15542v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.15542", "identifiers": {"arxiv": "2605.15542"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.15542"}], "name": "DRS-GUI: Dynamic Region Search for Training-Free GUI Grounding", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-05-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "DRS-GUI searches screenshot regions through focus, shift and scatter operations selected by a tree-search planner. A region-quality reward guides this training-free perceptual search before final GUI grounding.", "topics": ["grounding", "training-free", "planning", "screen-understanding"], "updated_at": "2026-05-15", "url": "https://arxiv.org/abs/2605.15542"}
{"aliases": [], "authors": ["Kean Shi", "Zihang Li", "Tianyi Ma", "Zengji Tu", "Jialong Wu", "Xinbo Xu", "Qingyao Yang", "Ruoyu Wu", "Weichu Xie", "Ming Wu", "Jason Zeng", "Michael Heinrich", "Elvis Zhang", "Liang Chen", "Kuan Li", "Baobao Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c4ee2439dd31a375b6f64cbc7f1d0ff75c87e1346b1627a53db145ae42682f8a", "submitted_at": "2026-05-15T09:35:15Z", "updated_at": "2026-05-24T16:35:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.15777v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "topics"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "8cfab6fca80090adb47c4ef634f89eedd49a6204dbdc2b5b8e0d6f04c32db289", "supports": ["summary", "environments", "scope", "notes"], "url": "https://raw.githubusercontent.com/UniPat-AI/SaaS-Bench/HEAD/README.md"}], "id": "arxiv:2605.15777", "identifiers": {"arxiv": "2605.15777"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.15777"}], "name": "SaaS-Bench: Can Computer-Use Agents Leverage Real-World SaaS to Solve Professional Workflows?", "notes": "The complete retained official abstract and cached official README were reviewed. The README establishes browser execution against running SaaS applications; no repository code was executed and benchmark results were not independently reproduced.", "published_at": "2026-05-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "SaaS-Bench evaluates browser agents on professional workflows across self-hosted applications. Weighted checkpoints measure task completion and partial progress, with application-state checks and configured content judges documented in the official harness.", "topics": ["web-navigation", "evaluation", "long-horizon", "workflow", "cross-app", "state-tracking"], "updated_at": "2026-05-24", "url": "https://arxiv.org/abs/2605.15777"}
{"aliases": [], "authors": ["Michael Solodko", "Justin Wagle"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "681ea16cca0a4bcc15a51144692babce934b6fc831cb3d845bc1ed5d3d48b264", "submitted_at": "2026-05-15T14:58:28Z", "updated_at": "2026-08-17T18:33:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.16024v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}], "id": "arxiv:2605.16024", "identifiers": {"arxiv": "2605.16024"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.16024"}], "name": "ScreenSearch: Uncertainty-Aware OS Exploration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-17", "url": "https://arxiv.org/abs/2605.16024"}
{"aliases": [], "authors": ["Chinmay Savadikar", "Mingyu Zhao", "Yuanzheng Zhu", "Han Li", "Shuang Xie", "Alberto Castelo", "Tianfu Wu", "Lingyun Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f5553ab9cf33840fefe0b5d3ec00229fd58898a98c09dc70902eca441c59fbd", "submitted_at": "2026-05-15T16:00:38Z", "updated_at": "2026-05-15T16:00:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16116v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.16116", "identifiers": {"arxiv": "2605.16116"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.16116"}], "name": "ShopGym: An Integrated Framework for Realistic Simulation and Scalable Benchmarking of E-Commerce Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-15", "url": "https://arxiv.org/abs/2605.16116"}
{"aliases": [], "authors": ["Ashwin Aravind"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "feeac1fa65eddc5e9d7e53c8ff492d2db2f4f5dfe05368221f917f151f05fa17", "submitted_at": "2026-03-24T11:39:35Z", "updated_at": "2026-03-24T11:39:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16265v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2605.16265", "identifiers": {"arxiv": "2605.16265"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.16265"}], "name": "AgentWall: A Runtime Safety Layer for Local AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-03-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-03-24", "url": "https://arxiv.org/abs/2605.16265"}
{"aliases": [], "authors": ["Haoren Zhao", "Tianyi Chen", "Zhen Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "671f33dd87a7ebf9233494ef6fa0efd9c9772a3951f3c218bf4de822825f7af2", "submitted_at": "2026-05-13T02:48:52Z", "updated_at": "2026-05-13T02:48:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16402v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.16402", "identifiers": {"arxiv": "2605.16402"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.16402"}], "name": "WinDeskGround: A Benchmark for Robust GUI Grounding in Complex Multi-Window Desktop Environments", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "WinDeskGround synthesizes multi-window desktop layouts with controlled occlusion, density and semantic similarity to test GUI grounding robustness. Its instruction-target benchmark measures how localization changes when interfaces depart from simple single-layer scenes.", "topics": ["grounding", "evaluation", "robustness", "data-generation"], "updated_at": "2026-05-13", "url": "https://arxiv.org/abs/2605.16402"}
{"aliases": [], "authors": ["Mike Wong", "Kevin Hsieh", "Suman Nath", "Ravi Netravali"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "660e6d4a10048411b7969bde7c55c8bb2b9521fe30892e09369196fd08846481", "submitted_at": "2026-05-15T19:12:43Z", "updated_at": "2026-05-19T13:45:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.16565v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.16565", "identifiers": {"arxiv": "2605.16565"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.16565"}], "name": "Skim: Speculative Execution for Fast and Efficient Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-19", "url": "https://arxiv.org/abs/2605.16565"}
{"aliases": [], "authors": ["Shilong Jin", "Lanjun Wang", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52c1bf4c614ac05a55f2c2c66d3ac4eeb1bf31743814a45eb9f0ce064b6baaab", "submitted_at": "2026-05-16T08:51:57Z", "updated_at": "2026-05-16T08:51:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16883v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.16883", "identifiers": {"arxiv": "2605.16883"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.16883"}], "name": "SE-GA: Memory-Augmented Self-Evolution for GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-05-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "SE-GA retrieves hierarchical memories during inference and uses the collected experience in a separate policy-improvement pipeline. Its evaluation includes offline grounding/action tasks and online Android interaction, which measure different capabilities.", "topics": ["memory", "interaction-history", "mobile-navigation", "grounding", "evaluation"], "updated_at": "2026-05-16", "url": "https://arxiv.org/abs/2605.16883"}
{"aliases": [], "authors": ["Zhiqiang Liu", "Wenhui Dong", "Yilang Tan", "Yuwen Qu", "Haochen Yin", "Chenyang Si"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd0ca5a262da8f4617c25ab3258d938dcf98147cde141dba94eb05b88c905188", "submitted_at": "2026-05-16T09:49:25Z", "updated_at": "2026-05-16T09:49:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16909v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.16909", "identifiers": {"arxiv": "2605.16909"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.16909"}], "name": "TOBench: A Task-Oriented Omni-Modal Benchmark for Real-World Tool-Using Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-16", "url": "https://arxiv.org/abs/2605.16909"}
{"aliases": [], "authors": ["Jihwan Kim", "Nikhil Parthasarathy", "Danfeng Qin", "Junhwa Hur", "Deqing Sun", "Bohyung Han", "Ming-Hsuan Yang", "Boqing Gong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b535d82bcf32f640c676c3b4d42a79be67beffa276dec4c361bd9d24ab569646", "submitted_at": "2026-05-17T05:02:52Z", "updated_at": "2026-05-23T08:41:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.17260v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.17260", "identifiers": {"arxiv": "2605.17260"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.17260"}], "name": "LiteFrame: Efficient Vision Encoders Unlock Frame Scaling in Video LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-23", "url": "https://arxiv.org/abs/2605.17260"}
{"aliases": [], "authors": ["Yutong Huang", "Vikranth Srivatsa", "Alex Asch", "Hansin Tushar Patwa", "Yiying Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51576fb9bde23f769d3200a0d593ad610eb27248917132aff1bcf6eb7ff03cea", "submitted_at": "2026-05-17T08:24:01Z", "updated_at": "2026-05-17T08:24:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.17320v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.17320", "identifiers": {"arxiv": "2605.17320"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.17320"}], "name": "TClone: Low-Latency Forking of Live GUI Environments for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "TClone versions live GUI workspaces through snapshots, isolated forks, rollback and selective commit or merge. It separates fast branch creation from durable checkpoints using copy-on-write sharing, filesystem versioning and asynchronous checkpointing, with end-to-end agent-loop latency measurements.", "topics": ["sandbox", "efficiency", "safety-security", "automation"], "updated_at": "2026-05-17", "url": "https://arxiv.org/abs/2605.17320"}
{"aliases": [], "authors": ["Udari Madhushani Sehwag", "Zhengyang Shan", "Heming Liu", "Dileepa Lakshan", "Joseph Brandifino", "Max Fenkell"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7692b9f3c3def47c872c0c1b8e720576861a90e27cacb3e61c966a9ed7098808", "submitted_at": "2026-05-17T08:30:45Z", "updated_at": "2026-05-17T08:30:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.17324v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.17324", "identifiers": {"arxiv": "2605.17324"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.17324"}], "name": "ASPI: Seeking Ambiguity Clarification Amplifies Prompt Injection Vulnerability in LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-17", "url": "https://arxiv.org/abs/2605.17324"}
{"aliases": [], "authors": ["Sirui Hong", "Zhijie Liu", "Tengfei Li", "Wei Tao", "Yifan Wu", "Chenglin Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e9167646e6433e6ebbcb8f1cc10c19b9b027b6e4f53fadf6df16751591a48369", "submitted_at": "2026-05-17T13:22:22Z", "updated_at": "2026-05-19T07:14:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.17439v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.17439", "identifiers": {"arxiv": "2605.17439"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.17439"}], "name": "DiagEval: Trajectory-Conditioned Diagnosis for Reliable Software Evaluation with GUI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-19", "url": "https://arxiv.org/abs/2605.17439"}
{"aliases": [], "authors": ["Guijin Son", "Jehyun Park", "Seyeon Park", "Sunghee Ahn", "Youngjae Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "247ae269abcab332f2e0c37fce310cbe50796c6f5dbe99110e1b54cfc2346870", "submitted_at": "2026-05-17T13:47:38Z", "updated_at": "2026-05-27T01:18:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.17448v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.17448", "identifiers": {"arxiv": "2605.17448"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.17448"}], "name": "Self-Improving CAD Generation Agents with Finite Element Analysis as Feedback", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2605.17448"}
{"aliases": [], "authors": ["Athar Parvez", "Muhammad Jawad Mufti", "Muqaddas Gull", "Omar Hammad"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff95e0b071508592190ba236741c1c01200dd38fdc74c31b2cc4d1bf8143ae9c", "submitted_at": "2026-05-17T21:14:32Z", "updated_at": "2026-05-17T21:14:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.17656v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.17656", "identifiers": {"arxiv": "2605.17656"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.17656"}], "name": "MUIAnno: An Expert-Annotated Dataset and Evaluation Benchmark for Mobile UI Understanding", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-05-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "MUIAnno provides expert labels for elements in manually explored mobile-app screens, with a web annotation tool that emits structured JSON. It also supplies UI-element detection baselines for the dataset.", "topics": ["screen-understanding", "grounding", "evaluation", "data-generation"], "updated_at": "2026-05-17", "url": "https://arxiv.org/abs/2605.17656"}
{"aliases": [], "authors": ["Yuxin Jin", "Siyuan Zhang", "Hanchen Wang", "Lu Qin", "Ying Zhang", "Wenjie Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65ceabc2925d6d597d2270b8a7cb5510e2f0e8ca6af70e8e595038fb6dc9556a", "submitted_at": "2026-05-18T00:50:23Z", "updated_at": "2026-05-18T00:50:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.17721v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.17721", "identifiers": {"arxiv": "2605.17721"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.17721"}], "name": "EXG: Self-Evolving Agents with Experience Graphs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-18", "url": "https://arxiv.org/abs/2605.17721"}
{"aliases": [], "authors": ["Jingjing Liu", "Ziye Huang", "Zihao Cheng", "Zeming Liu", "Jiahong Wu", "Yuhang Guo", "Kehai Chen", "Yunhong Wang", "Haifeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "017909a697c6c661d3486d960042f4d691f70a660a60083e70512de34690cb51", "submitted_at": "2026-05-18T08:36:04Z", "updated_at": "2026-05-18T08:36:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.18048v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.18048", "identifiers": {"arxiv": "2605.18048"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.18048"}], "name": "DocOS: Towards Proactive Document-Guided Actions in GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "DocOS benchmarks agents that search online documentation and translate procedural instructions into GUI actions in interactive web environments. It separates failures to retrieve useful procedures from failures to ground retrieved instructions into executable actions.", "topics": ["web-navigation", "grounding", "evaluation", "planning"], "updated_at": "2026-05-18", "url": "https://arxiv.org/abs/2605.18048"}
{"aliases": [], "authors": ["Ziyun Zeng", "Hang Hua", "Bocheng Zou", "Mu Cai", "Rogerio Feris", "Jiebo Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d96c8844274f79adea167169a0853e3a7cd82fee4a8a666913f97898a0836e7c", "submitted_at": "2026-05-18T16:57:36Z", "updated_at": "2026-05-18T16:57:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.18652v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.18652", "identifiers": {"arxiv": "2605.18652"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.18652"}], "name": "MementoGUI: Learning Agentic Multimodal Memory Control for Long-Horizon GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "MementoGUI learns a controller for selecting, compressing and retrieving textual and region-level visual memories while leaving the GUI-agent backbone fixed. It also converts computer-use traces into controller training data and evaluates semantic actions, progress and memory consistency on long-horizon tasks.", "topics": ["memory", "interaction-history", "long-horizon", "data-generation", "evaluation"], "updated_at": "2026-05-18", "url": "https://arxiv.org/abs/2605.18652"}
{"aliases": [], "authors": ["Felix Henry", "Lihan Lin", "Jiangyou Zhu", "Yangfan", "Bingqian Zhang", "Min Chen", "Shiyu Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "99e8e0039ec99ef8f6a6ef3900cb74a62ddd5ae4f8df30629110969290028e6a", "submitted_at": "2026-04-03T08:57:06Z", "updated_at": "2026-09-28T02:42:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.18758v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.18758", "identifiers": {"arxiv": "2605.18758"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.18758"}], "name": "OmniGUI: Benchmarking GUI Agents in Omni-Modal Smartphone Environments", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-04-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "OmniGUI evaluates step-level action prediction from interleaved smartphone screenshots, synchronous audio and video. Expert demonstrations carry multimodal-dependency annotations, and initial baselines use native multimodal models as agent proxies to study temporal and audio-related failures.", "topics": ["evaluation", "action-prediction", "video", "mobile-navigation"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2605.18758"}
{"aliases": [], "authors": ["Yuankai Li", "Tinghui Zhu", "Ha Min Son", "Zhe Zhao", "Xin Liu", "Muhao Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4fc83b6ed8bdaf4d2ab9512c41d7217c95973e6bac59a2c0892a51d4531dc0a", "submitted_at": "2026-05-19T02:13:29Z", "updated_at": "2026-05-19T02:13:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.19260v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.19260", "identifiers": {"arxiv": "2605.19260"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.19260"}], "name": "AQuaUI: Visual Token Reduction for GUI Agents with Adaptive Quadtrees", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-19", "url": "https://arxiv.org/abs/2605.19260"}
{"aliases": [], "authors": ["Xiaozhe Li", "Tianyi Lyu", "Yang Li", "Yichuan Ma", "Peiji Li", "Linyang Li", "Qipeng Guo", "Dahua Lin", "Kai Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "996f61455be8fc9839511d50e2930fe798c2355508d6d8ad0c73486d1a316967", "submitted_at": "2026-05-19T07:00:55Z", "updated_at": "2026-05-19T07:00:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.19447v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.19447", "identifiers": {"arxiv": "2605.19447"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.19447"}], "name": "What and When to Distill: Selective Hindsight Distillation for Multi-Turn Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-19", "url": "https://arxiv.org/abs/2605.19447"}
{"aliases": [], "authors": ["Haobo Hu", "Xiangwu Guo", "Zhiheng Chen", "Difei Gao", "Haotian Liu", "Libiao Jin", "Qi Mao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd78832d2550ce38e5147e8779455242a62f1aabdab652c676f7324c19578d4c", "submitted_at": "2026-05-19T07:35:22Z", "updated_at": "2026-05-19T07:35:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.19484v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.19484", "identifiers": {"arxiv": "2605.19484"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.19484"}], "name": "CutVerse: A Compositional GUI Agents Benchmark for Media Post-Production Editing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-19", "url": "https://arxiv.org/abs/2605.19484"}
{"aliases": [], "authors": ["Puyi Wang", "Yuhao Wang", "Linjie Li", "Zhengyuan Yang", "Kevin Qinghong Lin", "Yangguang Li", "Yu Cheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "23c22a0810194a505dca417d0bd3f7e60e9daac4d2c6fca46fdc45fd1b894fce", "submitted_at": "2026-05-19T09:31:04Z", "updated_at": "2026-05-19T09:31:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.19587v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.19587", "identifiers": {"arxiv": "2605.19587"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.19587"}], "name": "SceneCode: Executable World Programs for Editable Indoor Scenes with Articulated Objects", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-19", "url": "https://arxiv.org/abs/2605.19587"}
{"aliases": [], "authors": ["Jinbiao Wei", "Qianran Ma", "Yilun Zhao", "Xiao Zhou", "Kangqi Ni", "Guo Gan", "Arman Cohan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a93d880793018d1b47a39e0487e3be76550fd2b49de9535b2913e9f08159074", "submitted_at": "2026-05-19T12:40:29Z", "updated_at": "2026-05-19T12:40:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.19769v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.19769", "identifiers": {"arxiv": "2605.19769"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.19769"}], "name": "OpenComputer: Verifiable Software Worlds for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces OpenComputer, which builds desktop tasks around application-state verifiers and refines those verifiers using execution feedback. An evaluation harness records trajectories and assigns auditable partial credit across professional software workflows.", "topics": ["evaluation", "data-generation", "observability"], "updated_at": "2026-05-19", "url": "https://arxiv.org/abs/2605.19769"}
{"aliases": [], "authors": ["Fatemeh Pesaran Zadeh", "Seyeon Choi", "Xing Han Lù", "Siva Reddy", "Gunhee Kim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa6e11bae4d952d79276543ced6f81193b36868a6222fe51ee7b0816aa72fbab", "submitted_at": "2026-05-19T09:19:01Z", "updated_at": "2026-05-26T02:53:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.20291v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.20291", "identifiers": {"arxiv": "2605.20291"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.20291"}], "name": "Weasel: Out-of-Domain Generalization for Web Agents via Importance-Diversity Data Selection", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-05-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Weasel selects web-agent training steps by balancing importance and diversity across states, sites and interaction patterns. It also prunes training accessibility trees around known action targets and regenerates rationales to match the model’s reasoning style.", "topics": ["web-navigation", "imitation-learning", "efficiency", "robustness", "data-generation"], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2605.20291"}
{"aliases": [], "authors": ["Wujiang Xu", "Yu Wang", "Kai Mei", "Kaiqu Liang", "Zhenting Wang", "Mingyu Jin", "Han Zhang", "Shi-Xiong Zhang", "Wenyue Hua", "Sambit Sahu", "Dimitris N. Metaxas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89ee3bb59cd9293042d415b45ac4115f1931131a1f0ec443210c24614d9e337e", "submitted_at": "2026-05-20T07:25:33Z", "updated_at": "2026-05-20T07:25:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.20833v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2605.20833", "identifiers": {"arxiv": "2605.20833"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.20833"}], "name": "MemGym: a Long-Horizon Memory Environment for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-20", "url": "https://arxiv.org/abs/2605.20833"}
{"aliases": [], "authors": ["Caleb Winston", "Ron Yifeng Wang", "Azalia Mirhoseini", "Christos Kozyrakis"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b73bf7776e280a9040b1d7d1e29957913723253014c1dfe84f64916b2a08e5e", "submitted_at": "2026-05-20T17:54:27Z", "updated_at": "2026-05-29T17:54:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.21470v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.21470", "identifiers": {"arxiv": "2605.21470"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.21470"}], "name": "Agent JIT Compilation for Latency-Optimizing Web Agent Planning and Scheduling", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Agent JIT compilation turns browser task descriptions into validated executable plans containing model calls, tools and parallel work. A cost-based scheduler and pre/postcondition protocol target latency and incorrect tool use in web-agent execution.", "topics": ["planning", "efficiency", "tool-integration", "web-navigation"], "updated_at": "2026-05-29", "url": "https://arxiv.org/abs/2605.21470"}
{"aliases": [], "authors": ["Banghao Chi", "Yining Xie", "Mingyuan Wu", "Jingcheng Yang", "Jize Jiang", "Zhaoheng Li", "Shengyi Qian", "Minjia Zhang", "Klara Nahrstedt", "Rui Hou", "Xiangjun Fan", "Hanchao Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "22f8fbc30bb0aec89832ccb69c3762a019bf022aabd4232d8cd02b81287c743a", "submitted_at": "2026-05-21T15:47:41Z", "updated_at": "2026-05-21T15:47:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.22642v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.22642", "identifiers": {"arxiv": "2605.22642"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.22642"}], "name": "Spreadsheet-RL: Advancing Large Language Model Agents on Realistic Spreadsheet Tasks via Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-21", "url": "https://arxiv.org/abs/2605.22642"}
{"aliases": [], "authors": ["P. Ondratschek", "D. Przybylski", "H. N. Smitha", "J. Leenaarts", "R. Cameron", "S. K. Solanki"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5b8a033df037c77f764cfbea0e487b6c2be972752164d9e047df4e9b8f595170", "submitted_at": "2026-05-21T18:00:34Z", "updated_at": "2026-05-21T18:00:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.22916v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.22916", "identifiers": {"arxiv": "2605.22916"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.22916"}], "name": "Mg II h&k spectral line properties computed using 3D radiative transfer in an enhanced network region simulated with the MURaM-ChE code", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-21", "url": "https://arxiv.org/abs/2605.22916"}
{"aliases": [], "authors": ["Guiyao Tie", "Jiawen Shi", "Dingjie Song", "Yixiao Huang", "Ziji Sheng", "Xueyang Zhou", "Daizong Liu", "Pan Zhou", "Yongchao Chen", "Ran Xu", "Lifang He", "Qingsong Wen", "Manling Li", "Cong Lu", "Shuai Li", "Pengtao Xie", "Yixuan Yuan", "Rui Meng", "Lei Xing", "Lichao Sun", "Caiming Xiong", "Philip S. Yu", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6c32c3bd2974c26dfe7db2001984351da14ab7c84510c7f79c13736a44eec37", "submitted_at": "2026-05-22T03:40:30Z", "updated_at": "2026-05-22T03:40:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.23204v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.23204", "identifiers": {"arxiv": "2605.23204"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.23204"}], "name": "AutoResearch AI: Towards AI-Powered Research Automation for Scientific Discovery", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-22", "url": "https://arxiv.org/abs/2605.23204"}
{"aliases": [], "authors": ["Xirui Liu", "Sihang Zhou", "Yanning Hou", "Rong Zhou", "Haoyuan Chen", "Maolin He", "Siwei Wang", "Hao Chen", "Jian Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "afe80d68ecd01287a04d88f2afc64d4722677deefc5b670d80c188b4e16145c2", "submitted_at": "2026-04-28T11:39:20Z", "updated_at": "2026-04-28T11:39:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.23939v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.23939", "identifiers": {"arxiv": "2605.23939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.23939"}], "name": "DRIVE: Modeling Skills at the Reasoning and Interaction Levels for Web Agents under Continual Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-28", "url": "https://arxiv.org/abs/2605.23939"}
{"aliases": [], "authors": ["Yubo Li", "Yidi Miao", "Yuntian Shen", "Yuxin Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e64d36e920a33851aacf93138a2081d0470424dfec6a8fd11915928660824851", "submitted_at": "2026-05-24T00:07:25Z", "updated_at": "2026-05-26T04:25:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.24785v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2605.24785", "identifiers": {"arxiv": "2605.24785"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.24785"}], "name": "PANDO: Efficient Multimodal AI Agents via Online Skill Distillation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-05-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes inefficiency in VisualWebArena trajectories and distills reusable online skills to reduce repeated actions and token overhead in multimodal web agents.", "topics": ["efficiency", "memory", "web-navigation"], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2605.24785"}
{"aliases": [], "authors": ["Guohong Liu", "Jialei Ye", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Yunxin Liu", "Yuanchun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eb40c8bf005d56139d263e1b5dc4553ddcd68522124686b7f0865e6d8d1725ad", "submitted_at": "2026-05-24T16:33:14Z", "updated_at": "2026-06-18T02:24:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.25160v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.25160", "identifiers": {"arxiv": "2605.25160"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.25160"}], "name": "ScaleWoB: Guiding GUI Agents with Coding Agents via Large-Scale Environmental Synthesis", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. Mobile, desktop and automotive interfaces are simulated as webpages; this does not assert native deployment on those platforms.", "published_at": "2026-05-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ScaleWoB, which synthesizes interactive GUI environments as backend-free webpages with verifiable rewards. Its mobile-oriented benchmark evaluates agent execution and compares selected synthetic tasks with real-app behavior.", "topics": ["data-generation", "evaluation", "sandbox"], "updated_at": "2026-06-18", "url": "https://arxiv.org/abs/2605.25160"}
{"aliases": [], "authors": ["Andy Xu", "Yu-Wing Tai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1566e1fe5a87194166866eda81666a15d6c068a2612034eab948e58984ae17e6", "submitted_at": "2026-05-24T19:38:54Z", "updated_at": "2026-05-24T19:38:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.25233v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.25233", "identifiers": {"arxiv": "2605.25233"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.25233"}], "name": "Meta-Agent: From Task Descriptions to Verified Multi-Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-24", "url": "https://arxiv.org/abs/2605.25233"}
{"aliases": [], "authors": ["Bowen Wang", "Dunjie Lu", "Junli Wang", "Tianyi Bai", "Shixuan Liu", "Zhipeng Zhang", "Haiquan Wang", "Hao Hu", "Tianbao Xie", "Shuai Bai", "Dayiheng Liu", "Que Shen", "Junyang Lin", "Tao Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecc0712f55ee90c237f9a48e013e396b1082f932a485c22d8249d1da7acbf307", "submitted_at": "2026-05-25T09:28:03Z", "updated_at": "2026-06-08T08:32:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.25624v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.25624", "identifiers": {"arxiv": "2605.25624"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.25624"}], "name": "CUA-Gym: Scaling Verifiable Training Environments and Tasks for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The abstract promises future open sourcing; this annotation does not independently verify artifact release or licensing.", "published_at": "2026-05-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents CUA-Gym, which co-generates instructions, initial and target states, and executable reward checks through generator-discriminator iterations. Mock web applications expand training environments, with learned agents evaluated on OSWorld-Verified and WebArena.", "topics": ["data-generation", "reinforcement-learning", "sandbox"], "updated_at": "2026-06-08", "url": "https://arxiv.org/abs/2605.25624"}
{"aliases": [], "authors": ["Jingwei Sun", "Jianing Zhu", "Yuanyi Li", "Tongliang Liu", "Xia HU", "Bo Han"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d48aff278b0f77b0f28505ca16ffc121b5c10ab838b92c8c17254000b763c0ad", "submitted_at": "2026-05-25T11:09:22Z", "updated_at": "2026-05-25T11:09:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.25707v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.25707", "identifiers": {"arxiv": "2605.25707"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.25707"}], "name": "AgentHijack: Benchmarking Computer Use Agent Robustness to Common Environment Corruptions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-25", "url": "https://arxiv.org/abs/2605.25707"}
{"aliases": [], "authors": ["Cuong Huynh", "Maxim Popov", "Denis Gridusov", "Sergey Kolyubin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b4561bb5a3191606ed2cb9682a7ba98fc4ba2f20b75f70728909b3180a72601", "submitted_at": "2026-05-25T14:29:04Z", "updated_at": "2026-05-25T14:29:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.25901v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.25901", "identifiers": {"arxiv": "2605.25901"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.25901"}], "name": "AgentGrounder: Zero-Shot 3D Visual Pointcloud Grounding using Multimodal Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-25", "url": "https://arxiv.org/abs/2605.25901"}
{"aliases": [], "authors": ["Yusong Lin", "Xinyuan Liang", "Haiyang Wang", "Qipeng Gu", "Siqi Cheng", "Jiangui Chen", "Shuzhe Wu", "Feiyang Pan", "Lue Fan", "Sanyuan Zhao", "Dandan Tu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27999553928305cfa4223011feb3f9c2decda8a861e5800865537a81c7c08832", "submitted_at": "2026-05-25T17:50:04Z", "updated_at": "2026-05-25T17:50:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.26086v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.26086", "identifiers": {"arxiv": "2605.26086"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.26086"}], "name": "Claw-Anything: Benchmarking Always-On Personal Assistants with Broader Access to User's Digital World", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-05-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Claw-Anything benchmarks persistent personal assistance across activity histories, backend services and integrated GUI/CLI interaction. Its simulated event histories introduce irrelevant or conflicting context, and a separate generation pipeline supplies training environments.", "topics": ["evaluation", "interaction-history", "long-horizon", "robustness", "data-generation", "workflow"], "updated_at": "2026-05-25", "url": "https://arxiv.org/abs/2605.26086"}
{"aliases": [], "authors": ["Dingbang Wu", "Rui Hao", "Haiyang Wang", "Shuzhe Wu", "Han Xiao", "Zhenghong Li", "Bojiang Zhou", "Zheng Ju", "Zichen Liu", "Lue Fan", "Zhaoxiang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "145c7012fb15a85b203924ecd559c8b20d18b51bf79805a6751fec75d6f68dec", "submitted_at": "2026-05-25T17:59:49Z", "updated_at": "2026-09-23T17:59:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.26114v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.26114", "identifiers": {"arxiv": "2605.26114"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.26114"}], "name": "MobileGym: A Verifiable and Highly Parallel Simulation Platform for Mobile GUI Agent Research", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The simulator is browser-hosted; the abstract does not identify a native mobile operating system.", "published_at": "2026-05-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MobileGym, a browser-hosted simulation of mobile app interactions whose configurable JSON state supports deterministic judging and parallel rollouts. MobileGym-Bench adds parameterized tasks, with experiments examining online training and transfer to real devices.", "topics": ["sandbox", "evaluation", "reinforcement-learning"], "updated_at": "2026-09-23", "url": "https://arxiv.org/abs/2605.26114"}
{"aliases": [], "authors": ["JunJia Guo", "Yuhang Yao", "Chang Li", "Zimeng Pan", "Jiawei", "Zhou", "Jingdi Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64f8e09fc62f080a75aec1928b82187063cfc3ad8467290608347622410b6460", "submitted_at": "2026-05-22T20:29:12Z", "updated_at": "2026-09-29T00:12:17Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2605.26144v4"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.26144", "identifiers": {"arxiv": "2605.26144"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.26144"}], "name": "VISTA: An End-to-End Benchmark for Visual Spec-to-Web-App Coding Agents", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent GUI-application generation benchmark with real component interaction probes; its main evaluated agents are application coders, not general GUI task executors.", "published_at": "2026-05-22", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "VISTA evaluates coding agents that turn design handoffs into working web and Android applications. Its evaluator locates annotated interactive components and probes their behavior, keeping visual localization distinct from functional correctness.", "topics": ["evaluation", "ui-design", "automation"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2605.26144"}
{"aliases": [], "authors": ["Abdelghny Orogat", "Essam Mansour"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5639dabccb22430c1a536045b04069a2e67bf4538d4c845341736a2d0bda9c7c", "submitted_at": "2026-05-25T18:22:42Z", "updated_at": "2026-05-25T18:22:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.26252v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.26252", "identifiers": {"arxiv": "2605.26252"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.26252"}], "name": "Is Agent Memory a Database? Rethinking Data Foundations for Long-Term AI Agent Memory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-25", "url": "https://arxiv.org/abs/2605.26252"}
{"aliases": [], "authors": ["Runxi Huang", "Liyu Zhang", "Shengzhong Liu", "Xiaomin Ouyang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97467a9a1d3c456a0078e96ebb128efa43412a3139a7a68beb4e6dd7b5f0674d", "submitted_at": "2026-05-26T04:53:53Z", "updated_at": "2026-05-26T04:53:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.26546v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.26546", "identifiers": {"arxiv": "2605.26546"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.26546"}], "name": "MobileExplorer: Accelerating On-Device Inference for Mobile GUI Agents via Online Exploration", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-05-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileExplorer probes relevant UI elements while an on-device model is reasoning, then supplies the exploration trace as compact context. A two-level rollback mechanism restores the working state when simple backtracking fails.", "topics": ["mobile-navigation", "efficiency", "memory", "robustness"], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2605.26546"}
{"aliases": [], "authors": ["Yiqun Chen", "Wei Yang", "Erhan Zhang", "Shijie Wang", "Qi Liu", "Zechun Niu", "Bin Zhang", "Haitao Li", "Rui Li", "Lingyong Yan", "Jinyuan Feng", "Biqing Qi", "Xiaochi Wei", "Yan Gao", "Yi Wu", "Yao Hu", "Jiaxin Mao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5206bc6d63599b8f4b4fd7020dd3bab43270fc0ab037f750d9d47ce74b9812c6", "submitted_at": "2026-05-26T07:30:03Z", "updated_at": "2026-05-26T07:30:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.26646v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.26646", "identifiers": {"arxiv": "2605.26646"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.26646"}], "name": "UnityMAS-O: A General RL Optimization Framework for LLM-Based Multi-Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2605.26646"}
{"aliases": [], "authors": ["Heng Qu", "Yike Liu", "Renren Jin", "Wenzong Zhang", "Pengzhi Gao", "Wei Liu", "Jian Luan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c12f0337cbc49b1dbb7e4192a5af1a1704d4c561712118ab92147ba09ea8d880", "submitted_at": "2026-05-26T15:03:56Z", "updated_at": "2026-05-26T15:03:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.27134v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2605.27134", "identifiers": {"arxiv": "2605.27134"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.27134"}], "name": "Scaling, Benchmarking, and Reasoning of Vision-Language Agents for Mobile GUI Navigation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2605.27134"}
{"aliases": [], "authors": ["Yuxin Chen", "Xiaodong Cai", "Junfeng Fang", "Zhuowen Han", "Yu Wang", "Yaorui Shi", "Yi Zhang", "Qi Gu", "Xunliang Cai", "Xiang Wang", "An Zhang", "Tat-Seng Chua"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ae5e3ec47891f06acdd2eeb269110a9c75704f8037d0c0fd76b0f363c13bc03", "submitted_at": "2026-05-26T16:02:00Z", "updated_at": "2026-05-26T16:02:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.27209v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.27209", "identifiers": {"arxiv": "2605.27209"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.27209"}], "name": "Learning to Act under Noise: Enhancing Agent Robustness via Noisy Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2605.27209"}
{"aliases": [], "authors": ["Zongheng Cao", "Yi Zheng", "Rui Song", "Xinyu Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3b026cff7b10428a3fd2a6023e9e492a28c5f893a8730ca9ab3a8b79b20571a", "submitted_at": "2026-05-26T21:27:16Z", "updated_at": "2026-05-26T21:27:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.27705v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.27705", "identifiers": {"arxiv": "2605.27705"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.27705"}], "name": "AgenticVBench: Can AI Agents Complete Real-World Post-Production Tasks?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2605.27705"}
{"aliases": [], "authors": ["Yifan Sui", "Xin Huang", "Hongbing Li", "Fang Xu", "Jiahe Lv", "Haolong Yan", "Yeqing Shen", "Litao Liu", "Zhimin Fan", "Ziyang Meng", "Jia Wang", "Junbo Qi", "Kaijun Tan", "Zheng Ge", "Xiangyu Zhang", "Daxin Jiang", "Osamu Yoshie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0c89b8b5ef5aa17fe747201285104544648449804cc2229a77577d217763f497", "submitted_at": "2026-05-26T23:19:42Z", "updated_at": "2026-05-26T23:19:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.27761v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.27761", "identifiers": {"arxiv": "2605.27761"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.27761"}], "name": "AndroidDaily: A Verifiable Benchmark for Mobile GUI Agents on Real-World Closed-Source Applications", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "AndroidDaily evaluates mobile GUI agents on daily workflows in closed-source Android apps. Its GRADE evaluator follows visual trajectories against operational obligations, output-quality requirements and negative constraints, producing process-level diagnostics without hidden app state.", "topics": ["mobile-navigation", "evaluation", "long-horizon", "robustness"], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2605.27761"}
{"aliases": [], "authors": ["Pengyu Zhu", "Lijun Li", "Yaxing Lyu", "Qianxin Luo", "Jingyi Yang", "Yi Liu", "Tingfeng Hui", "Xinyu Yuan", "Li Sun", "Sen Su", "Jing Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e0ffbe367e6c4cd8c9a08a23d849187985b646222286bb876d8fab3491a627e", "submitted_at": "2026-05-27T03:20:45Z", "updated_at": "2026-09-01T11:09:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.27898v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.27898", "identifiers": {"arxiv": "2605.27898"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.27898"}], "name": "UniACE: A Unified Framework for Evaluating LLM Agentic Capabilities", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-01", "url": "https://arxiv.org/abs/2605.27898"}
{"aliases": [], "authors": ["Ruoqi Guo", "Yi Liu", "Gelei Deng", "Yiheng Xiong", "Yuekang Li", "Ying Zhang", "Leo Yu Zhang", "Lida Zhao", "Ji Jie", "Yuxiao Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4f39433695b9574b275fc844ca6b10a43baf66aed57048bf841f41bf08d4ac9a", "submitted_at": "2026-05-27T08:09:53Z", "updated_at": "2026-05-27T08:09:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.28116v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.28116", "identifiers": {"arxiv": "2605.28116"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.28116"}], "name": "MIRAGE: Context-Aware Prompt Injection against Mobile GUI Agents via User-Generated Content", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "MIRAGE constructs mobile screenshot injections in ordinary user-content regions through localization, contextual payload generation and realism curation. Its evaluation separates visual realism from attack effectiveness against pixel-observing GUI agents.", "topics": ["safety-security", "data-generation", "evaluation", "robustness"], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2605.28116"}
{"aliases": [], "authors": ["Chenyu Zhou", "Xinyun Lu", "Jiangyue Zhao", "Jianghao Lin", "Dongdong Ge", "Yinyu Ye"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15e04680cb367e9bb410558f3ebe0df24f7a52762322ce5dc58a1719de48be76", "submitted_at": "2026-05-27T08:41:30Z", "updated_at": "2026-05-27T08:41:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.28158v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.28158", "identifiers": {"arxiv": "2605.28158"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.28158"}], "name": "OR-Space: A Full-Lifecycle Workspace Benchmark for Industrial Optimization Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2605.28158"}
{"aliases": [], "authors": ["Maharshi Gor", "Yoo Yeon Sung", "Yu Hou", "Eve Fleisig", "Irene Ying", "Tianyi Zhou", "Jordan Boyd-Graber"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e0a5588aeaf0df8e5011eb59ab64a06222e842348544d2cef62d260293139911", "submitted_at": "2026-05-27T10:05:16Z", "updated_at": "2026-05-27T10:05:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.28255v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.28255", "identifiers": {"arxiv": "2605.28255"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.28255"}], "name": "AI, Take the Wheel: What Drives Delegation and Trust in Human-Computer Cooperative Question Answering?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2605.28255"}
{"aliases": [], "authors": ["Yixu Huang", "Bo Li", "Na Li", "Zhe Wang", "Kaijie Chen", "Haonan Ge", "Qingyi Si", "Yuanzhe Shen", "Ruihan Yang", "Guangjing Wang", "Hongcheng Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4e03e0ef06c6610eaefeec08b7891a1ab4d5073c0793650d4845735259978723", "submitted_at": "2026-05-27T10:08:48Z", "updated_at": "2026-10-02T05:26:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.28258v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.28258", "identifiers": {"arxiv": "2605.28258"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.28258"}], "name": "GUI Agents for Continual Game Generation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced. GUI agents actually load and play browser-game builds, providing an executable interface-testing signal for game-code generation; the central application remains game development.", "published_at": "2026-05-27", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "PlaytestArena evaluates generated browser games using expected in-play behaviors adjudicated by GUI agents. Play2Code pairs a coding agent with a rubric-blind GUI playtester whose interaction traces and feedback guide iterative refinement, with a separate final judge.", "topics": ["evaluation", "automation", "workflow"], "updated_at": "2026-10-02", "url": "https://arxiv.org/abs/2605.28258"}
{"aliases": [], "authors": ["Zheng Wu", "Chengcheng Han", "Zhengxi Lu", "Tianjie Ju", "Yanyu Chen", "Qi Gu", "Xunliang Cai", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "78a5f5d7df9a8b6f1dbdb5693849c728a4b97e1f25c6e2e24d0c2418628c64e7", "submitted_at": "2026-05-27T14:29:41Z", "updated_at": "2026-05-27T14:29:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.28534v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:11:36Z", "role": "primary", "sha256": "235f1fff1385719d603377b680d74f6803bbe46d112dd6c008bee6b42f105429", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.28534", "identifiers": {"arxiv": "2605.28534"}, "kind": "paper", "last_verified_at": "2026-10-17T17:11:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.28534"}], "name": "GUI-CIDER: Mid-training GUI Agents via Causal Internalization and Density-aware Exemplar Reselection", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-CIDER turns GUI trajectories into textual planning and causal interface knowledge, reselects exemplars for causal structure and low redundancy, then mid-trains GUI agents on that corpus.", "topics": ["data-generation", "planning"], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2605.28534"}
{"aliases": [], "authors": ["Yanqiu Zhao", "Dongying Zheng", "Kaibo Huang", "Yukun Wei", "Zhongliang Yang", "Linna Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6eb9b67f7e51edecfbe57af47b6bad445c05129c11c377d32ea84369b430f895", "submitted_at": "2026-05-27T15:51:22Z", "updated_at": "2026-05-29T04:03:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.28646v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.28646", "identifiers": {"arxiv": "2605.28646"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.28646"}], "name": "MaskClaw: Edge-Side Personalized Privacy Arbitration for GUI Agents with Behavior-Driven Skill Evolution", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-05-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "MaskClaw makes local Allow, Mask or Ask decisions before GUI screenshots leave a trusted environment, using visual evidence and personalized policy memory. Its benchmark and designed evolution scenarios test how corrections and edits can become privacy skills checked by a sandbox gate.", "topics": ["safety-security", "memory", "human-in-the-loop", "evaluation"], "updated_at": "2026-05-29", "url": "https://arxiv.org/abs/2605.28646"}
{"aliases": [], "authors": ["Suji Kim", "Kangsan Kim", "Sung Ju Hwang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7efd7a912c16dabc9e9bcf24752ce526e639ea34b52be1989aae544fc23af9f9", "submitted_at": "2026-05-27T17:37:00Z", "updated_at": "2026-05-27T17:37:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.28775v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.28775", "identifiers": {"arxiv": "2605.28775"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.28775"}], "name": "Learn from Weaknesses: Automated Domain Specialization for Small Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2605.28775"}
{"aliases": [], "authors": ["Yifei He", "Rui Yang", "Hao Bai", "Tong Zhang", "Han Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af13a82ab8469a5be0b5500f067cb32cb7ecb54ec15f7178c736d6470374def8", "submitted_at": "2026-05-27T21:28:26Z", "updated_at": "2026-05-27T21:28:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29119v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.29119", "identifiers": {"arxiv": "2605.29119"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.29119"}], "name": "PRO-CUA: Process-Reward Optimization for Computer Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-27", "url": "https://arxiv.org/abs/2605.29119"}
{"aliases": [], "authors": ["Tenghao Huang", "Kung-Hsiang Huang", "Prafulla Kumar Choubey", "Yilun Zhou", "Muhao Chen", "Jonathan May", "Chien-Sheng Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f756d85a19896642fe8f2f4aecd24fac486715c012ba2196cf1a7fcac6c0d76", "submitted_at": "2026-05-28T01:05:50Z", "updated_at": "2026-05-28T01:05:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29218v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.29218", "identifiers": {"arxiv": "2605.29218"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.29218"}], "name": "GTA: Generating Long-Horizon Tasks for Web Agents at Scale", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.29218"}
{"aliases": [], "authors": ["Junyang Wang", "Haiyang Xu", "Xi Zhang", "Zhaoqing Zhu", "Ming Yan", "Jieping Ye", "Jitao Sang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fdef1adbf2ae7142ffee8cd40906e5fbfc7903ef84b688ae891f7ddf5e75d228", "submitted_at": "2026-05-28T04:00:13Z", "updated_at": "2026-05-28T04:00:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29324v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.29324", "identifiers": {"arxiv": "2605.29324"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.29324"}], "name": "STAMP: Training Explicit Memory for Mobile GUI Agents in Controllable and Scalable Virtual Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.29324"}
{"aliases": [], "authors": ["Chengzhi Liu", "Yuzhe Yang", "Sophia Xiao Pu", "Yepeng Liu", "Lin Long", "Yichen Guo", "Nuo Chen", "Zhaotian Weng", "Elena Kochkina", "Simerjot Kaur", "Charese Smiley", "Xiaomo Liu", "James Zou", "Sheng Liu", "Yuheng Bu", "Songyou Peng", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bc2247df361d1a2e929c222aa81c40a29c8d193e902ea7d25cdd9e28418c4cb4", "submitted_at": "2026-05-28T04:27:20Z", "updated_at": "2026-06-01T07:01:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.29341v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.29341", "identifiers": {"arxiv": "2605.29341"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.29341"}], "name": "WorldMemArena: Evaluating Multimodal Agent Memory Through Action-World Interaction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-01", "url": "https://arxiv.org/abs/2605.29341"}
{"aliases": [], "authors": ["Masafumi Enomoto", "Ryoma Obara", "Haochen Zhang", "Masafumi Oyamada"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0f765cdf6adacc06f55d3a279cce986ffd12da8c71dd6a2e486bf39ed151350", "submitted_at": "2026-05-28T05:46:39Z", "updated_at": "2026-05-28T05:46:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29397v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.29397", "identifiers": {"arxiv": "2605.29397"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.29397"}], "name": "Revisiting Observation Reduction for Web Agents: Comprehensive Evaluation with a Lightweight Framework", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.29397"}
{"aliases": [], "authors": ["Tianpeng Bu", "Xin Liu", "Qihua Chen", "Hao Jiang", "Shurui Li", "Hongtao Duan", "Lu Jiang", "Lulu Hu", "Bin Yang", "Minying Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0b6e406ae5fd009cd0602360276a7df951977170840b8560c8ca10136d838848", "submitted_at": "2026-05-28T06:40:24Z", "updated_at": "2026-05-28T06:40:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29447v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.29447", "identifiers": {"arxiv": "2605.29447"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.29447"}], "name": "Recovering Policy-Induced Errors: Benchmarking and Trajectory Synthesis for Robust GUI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.29447"}
{"aliases": [], "authors": ["Prakhar Dixit", "Sadia Kamal", "Tim Oates"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e206e12d819482db6c21e58187b93e4401055a04066cc2ffc5d3423973d08add", "submitted_at": "2026-05-28T06:56:42Z", "updated_at": "2026-05-31T05:31:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.29463v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.29463", "identifiers": {"arxiv": "2605.29463"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.29463"}], "name": "Honest Lying: Understanding Memory Confabulation in Reflexive Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-31", "url": "https://arxiv.org/abs/2605.29463"}
{"aliases": [], "authors": ["Yuxuan Liu", "Xin Lai", "Junyi Li", "Pengyuan Lyu", "Jason", "Yiduo Guo", "Zhengyao Fang", "Yang Ding", "Yi Zhang", "Weinong Wang", "Huawen Shen", "Xingran Zhou", "Liang Wu", "Fei Tang", "Sunqi Fan", "Shangpin Peng", "Zheng Ruan", "Anran Zhang", "Chengquan Zhang", "Han Hu", "Benyou Wang", "Ji-Rong Wen", "Rui Yan", "Zhengyang Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a41cb1a93e6aa6ba8fbc7dca55415577ab5880a6592e9b0bee5b12b5a7732cbf", "submitted_at": "2026-05-28T07:14:15Z", "updated_at": "2026-09-28T15:27:12Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.29486v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}], "id": "arxiv:2605.29486", "identifiers": {"arxiv": "2605.29486"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.29486"}], "name": "PhoneWorld: From Real-App Trajectories to Dynamic and Verifiable Environments for Phone-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.", "published_at": "2026-05-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PhoneWorld, which converts observed app pages, transitions and state changes into executable Android environments through a build-inspect-repair process. Generated tasks and programmatic verifiers support benchmarking, trajectory collection and online training.", "topics": ["data-generation", "sandbox", "reinforcement-learning", "evaluation"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2605.29486"}
{"aliases": [], "authors": ["Yuxiang Chai", "Han Xiao", "Xinyu Fu", "Jinpeng Chen", "Rui Liu", "Hongsheng Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "461790f351a2816004027cc58d07f79ee7c670ed5064c59b1651a4efb84ff2bf", "submitted_at": "2026-05-28T07:49:09Z", "updated_at": "2026-05-28T07:49:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29534v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.29534", "identifiers": {"arxiv": "2605.29534"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.29534"}], "name": "UI-KOBE: Knowledge-Oriented Behavior Exploration for Lightweight Graph-Guided GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-05-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "UI-KOBE explores mobile applications to build graphs of UI states and executable transitions. A lightweight agent then matches its current screenshot to the graph and chooses a transition, local action, completion signal or fallback action.", "topics": ["mobile-navigation", "planning", "state-tracking", "memory", "efficiency"], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.29534"}
{"aliases": [], "authors": ["Linhao Zhang", "Aiwei Liu", "Yuan Liu", "Xiao Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b56124d77060b30f678642bc046d4295822255bf14b4087659be403da01f2f25", "submitted_at": "2026-05-28T08:50:34Z", "updated_at": "2026-05-28T08:50:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29615v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}], "id": "arxiv:2605.29615", "identifiers": {"arxiv": "2605.29615"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.29615"}], "name": "DiffSpot: Can VLMs Spot Fine-Grained Visual Differences in Web Interfaces?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.29615"}
{"aliases": [], "authors": ["Dongrui Liu", "Yu Li", "Zhonghao Yang", "Peng Wang", "Guanxu Chen", "Yuejin Xie", "Qinghua Mao", "Wanying Qu", "Yanxu Zhu", "Tianyi Zhou", "Leitao Yuan", "Zhijie Zheng", "Qihao Lin", "Yimin Wang", "Haoyu Luo", "Shuai Shao", "Chen Qian", "Qingyu Liu", "Ling Tang", "Ruiyang Qin", "Qihan Ren", "Junxiao Yang", "Kun Wang", "Zhiheng Xi", "Linfeng Zhang", "Ranjie Duan", "Bo Zhang", "Wenjie Wang", "Wen Shen", "Qiaosheng Zhang", "Yan Teng", "Chaochao Lu", "Rui Mei", "Man Li", "Jialing Tao", "Xi Lin", "Tianhang Zheng", "Yong Liu", "Quanshi Zhang", "Lei Zhu", "Xingjun Ma", "Junhua Liu", "Hui Xue", "Xiaoxiang Zuo", "Xiangnan He", "Chao Shen", "Xianglong Liu", "Minlie Huang", "Jing Shao", "Xia Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9f18a697aa8aba3d02b9ac99ee952f033b6e116ec0186d33619b58d14a223f29", "submitted_at": "2026-05-28T11:48:37Z", "updated_at": "2026-05-28T11:48:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29801v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.29801", "identifiers": {"arxiv": "2605.29801"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.29801"}], "name": "AgentDoG 1.5: A Lightweight and Scalable Alignment Framework for AI Agent Safety and Security", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.29801"}
{"aliases": [], "authors": ["Alejandra Zambrano", "Sara Vera Marjanovic", "Imene Kerboua", "Xing Han Lù", "Leila Kosseim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ce309f6b2b4bce74fe5991599b7117ee7d0d89645987f99a89844b93d93ce6f", "submitted_at": "2026-05-28T13:39:49Z", "updated_at": "2026-05-28T13:39:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29927v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.29927", "identifiers": {"arxiv": "2605.29927"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.29927"}], "name": "Does The Way You Plan Matter? An Empirical Study of Planning Representations for LLM Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.29927"}
{"aliases": [], "authors": ["Hesong Wang", "Xin Jin", "Lu Lu", "Chenhaowen Li", "Jian Chen", "Qiang Liu", "Huan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5d068cae3c5a60be26fb36127c972c5eae4d5eb71537e3c8c80be80672cb6a7e", "submitted_at": "2026-05-28T14:36:44Z", "updated_at": "2026-05-28T14:36:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.30010v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.30010", "identifiers": {"arxiv": "2605.30010"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.30010"}], "name": "EarlyTom: Early Token Compression Completes Fast Video Understanding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.30010"}
{"aliases": [], "authors": ["Parsa Mazaheri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "beb170ac91a87fa3cc283c626f5af2caa2383ba67d94eabb3e2e362e4a0fffd3", "submitted_at": "2026-05-28T15:03:17Z", "updated_at": "2026-05-28T15:03:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.30052v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.30052", "identifiers": {"arxiv": "2605.30052"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.30052"}], "name": "REPOT: Recoverable Program-of-Thought via Checkpoint Repair", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.30052"}
{"aliases": [], "authors": ["Selim Kuzucu", "Alessio Tonioni", "Vasile Lup", "Bernt Schiele", "Federico Tombari", "Muhammad Ferjad Naeem"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d64d3b12d4616440101ea9839b3abe2ada4b5ec0113e0fbd67105e7552bef651", "submitted_at": "2026-05-28T15:57:31Z", "updated_at": "2026-05-28T15:57:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.30126v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.30126", "identifiers": {"arxiv": "2605.30126"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.30126"}], "name": "PARCEL: Pool-Anchored Resampling with Conditioned Elastic Queries for Efficient Vision-Language Understanding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2605.30126"}
{"aliases": [], "authors": ["Kewei Xu", "Xiaoben Lu", "Shuofei Qiao", "Zihan Ding", "Haoming Xu", "Lei Liang", "Ningyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "697bf203870061e183ccf38f71cc9df0b4396ddf6807cac6d882fb7ea7f692c6", "submitted_at": "2026-05-28T18:00:20Z", "updated_at": "2026-09-05T15:19:36Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.30434v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.30434", "identifiers": {"arxiv": "2605.30434"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.30434"}], "name": "LongDS-Bench: On the Failure of Long-Horizon Agentic Data Analysis", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-05", "url": "https://arxiv.org/abs/2605.30434"}
{"aliases": [], "authors": ["Junlong Li", "Chao Hao", "Lap-Pui Chau", "Yi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dca6a268882b5e740d88188c625f3587f9c8c18b62a8e302428680e68f0f3978", "submitted_at": "2026-05-29T06:17:53Z", "updated_at": "2026-05-29T06:17:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.30884v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.30884", "identifiers": {"arxiv": "2605.30884"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.30884"}], "name": "GUI-C$^2$: Coarse-to-Fine GUI Grounding via Difficulty-Aware Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-29", "url": "https://arxiv.org/abs/2605.30884"}
{"aliases": [], "authors": ["Srivatsa Kundurthy", "Clara Na", "Colton Moraine", "Anoushka Mohta", "Case Winter", "George Fang", "John Ling", "Emma Strubell", "Zach Kirshner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60fcacb09c856c24b1afb35ff426e6bae0d03b02086333389cc301f270fd53d0", "submitted_at": "2026-05-29T06:43:23Z", "updated_at": "2026-05-29T06:43:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.30907v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.30907", "identifiers": {"arxiv": "2605.30907"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.30907"}], "name": "BlueFin: Benchmarking LLM Agents on Financial Spreadsheets", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-29", "url": "https://arxiv.org/abs/2605.30907"}
{"aliases": [], "authors": ["Tianyi Zhou", "Dongrui Liu", "Leitao Yuan", "Jing Shao", "Xia Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "80034a42d597be3f746bc9698e46697003002946b997da4350ad277110151472", "submitted_at": "2026-05-29T12:59:08Z", "updated_at": "2026-05-29T12:59:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.31264v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.31264", "identifiers": {"arxiv": "2605.31264"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.31264"}], "name": "COLLEAGUE.SKILL: Automated AI Skill Generation via Expert Knowledge Distillation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-29", "url": "https://arxiv.org/abs/2605.31264"}
{"aliases": [], "authors": ["Weile Chen", "Bingchen Miao", "Qifan Yu", "Wendong Bu", "Guoming Wang", "Wenqiao Zhang", "Shengyu Zhang", "Juncheng Li", "Siliang Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "46c624e47b943201cceba50f85331506d6044522d4a3b3ef876553d1d29a8e83", "submitted_at": "2026-05-29T14:37:27Z", "updated_at": "2026-05-29T14:37:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.31365v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2605.31365", "identifiers": {"arxiv": "2605.31365"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.31365"}], "name": "Learning to Adapt: Self-Improving Web Agent via Cognitive-Aware Exploration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-29", "url": "https://arxiv.org/abs/2605.31365"}
{"aliases": [], "authors": ["Rosario Forte", "Giuseppe Lando", "Antonino Furnari"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cd5a3c7ce91269c42d26b7c02a445474ccaf5435a643f7b93a8fcaa0ba723655", "submitted_at": "2026-05-29T17:20:10Z", "updated_at": "2026-06-01T11:50:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.31557v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2605.31557", "identifiers": {"arxiv": "2605.31557"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2605.31557"}], "name": "EGOSTREAM: A Diagnostic Benchmark for Streaming Episodic Memory in Egocentric Vision", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-01", "url": "https://arxiv.org/abs/2605.31557"}
{"aliases": [], "authors": ["Neil F. Johnson", "Ferney J. Rodriguez", "Luis Quiroga"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7189466bcee61c41c106c01f6118090d084a9d759482ed475cb9c334f949c10a", "submitted_at": "2026-05-29T18:41:16Z", "updated_at": "2026-05-29T18:41:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.00254v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.00254", "identifiers": {"arxiv": "2606.00254"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.00254"}], "name": "Symmetry-Protected Quantum Computing using Metamaterials", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-29", "url": "https://arxiv.org/abs/2606.00254"}
{"aliases": [], "authors": ["Jeremy Tien", "Abishek Anand", "Yu-Rou Tuan", "Yuchen Shen", "J. Zico Kolter", "Aran Nayebi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "469c3b75b7dc63318d7ba777d9ae522d71247afdccc3d8cfd723a8780bcbde76", "submitted_at": "2026-05-29T20:29:35Z", "updated_at": "2026-10-01T15:01:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.00341v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.00341", "identifiers": {"arxiv": "2606.00341"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.00341"}], "name": "ROGUE: Evaluating Corrigibility Failures in Frontier Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2606.00341"}
{"aliases": [], "authors": ["Soham Roy", "Sarthakbrata Halder", "Arya Bharaty", "Vaibhav Bhaskar", "Yash Sinha", "Dhruv Kumar", "Srikant Panda", "Murari Mandal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a17c8c0e0aa0f1a314472a0530c898bbf562a5b6bd9bae82615cc826796d0117", "submitted_at": "2026-05-30T03:00:28Z", "updated_at": "2026-05-30T03:00:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.00497v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.00497", "identifiers": {"arxiv": "2606.00497"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.00497"}], "name": "\"I Strongly Suspect This Website Is a Scam\": Benchmarking PII Leakage and Detection without Defense in Autonomous Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-30", "url": "https://arxiv.org/abs/2606.00497"}
{"aliases": [], "authors": ["Tyler Stennett", "Rangeet Pan", "Bridget McGinn", "Alessandro Orso", "Saurabh Sinha"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc2053b253b86c393a6828b72291524d483fc11b2db3724f025c827a0772a686", "submitted_at": "2026-05-30T04:49:10Z", "updated_at": "2026-08-13T19:42:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.00530v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.00530", "identifiers": {"arxiv": "2606.00530"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.00530"}], "name": "Sakura: An Approach for Generating Complex Tests from Natural Language Test Descriptions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-13", "url": "https://arxiv.org/abs/2606.00530"}
{"aliases": [], "authors": ["Yiling Gao", "Hongchen Wei", "Zhenzhong Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "32ab60785c110ddd8db1883739303f7212380a993d955aefeb8364537eb6a9f0", "submitted_at": "2026-05-30T05:28:43Z", "updated_at": "2026-07-23T02:38:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.00543v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.00543", "identifiers": {"arxiv": "2606.00543"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.00543"}], "name": "ETC: Extreme Token Compression via Task-aware Visual Information Distillation in VLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-23", "url": "https://arxiv.org/abs/2606.00543"}
{"aliases": [], "authors": ["Alexandre Dulaunoy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a08b7828bdf50ba9978838e1751af28445b72c01870cd75361f0260e85dc900e", "submitted_at": "2026-05-30T19:03:39Z", "updated_at": "2026-05-30T19:03:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.00856v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.00856", "identifiers": {"arxiv": "2606.00856"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.00856"}], "name": "GCVE: A Decentralized Model for Vulnerability Identification, Publication, and Operational Enrichment", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-30", "url": "https://arxiv.org/abs/2606.00856"}
{"aliases": [], "authors": ["Yipeng Gao", "Lei Shu", "Genzhi Ye", "Xi Xiong", "Ameesh Makadia", "Meiqi Guo", "Laurent Itti", "Jindong Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f4293c2206665da268166e6ce49c4f66d52007b80ef027e666350e4fb91a0b56", "submitted_at": "2026-05-31T06:59:49Z", "updated_at": "2026-05-31T06:59:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.01057v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.01057", "identifiers": {"arxiv": "2606.01057"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.01057"}], "name": "3DCodeBench: Benchmarking Agentic Procedural 3D Modeling Via Code", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-05-31", "url": "https://arxiv.org/abs/2606.01057"}
{"aliases": [], "authors": ["Yunhao Feng", "Xiaohu Du", "Xinhao Deng", "Yifan Ding", "Ming Wen", "Yixu Wang", "Yuxiang Xie", "Baihui Zheng", "Yingshui Tan", "Yige Li", "Yutao Wu", "Kerui Cao", "Wenke Huang", "Yanming Guo", "Xingjun Ma", "Yu-Gang Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea8b37a42d6701a09fed0edfa5871b1b1d5a1ca2532db3571cd8d306983cd985", "submitted_at": "2026-05-31T11:16:18Z", "updated_at": "2026-06-02T05:59:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.01166v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.01166", "identifiers": {"arxiv": "2606.01166"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.01166"}], "name": "BraveGuard: From Open-World Threats to Safer Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-02", "url": "https://arxiv.org/abs/2606.01166"}
{"aliases": [], "authors": ["Jing Yu Koh", "Ruslan Salakhutdinov", "Daniel Fried"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "acc6e8af676b7de5393469d379746e674ccd662f570399622d0d5b1b0e882939", "submitted_at": "2026-06-01T01:29:36Z", "updated_at": "2026-06-01T01:29:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.01533v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.01533", "identifiers": {"arxiv": "2606.01533"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.01533"}], "name": "Multi-Agent Computer Use", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "MACU represents computer-use objectives as a dependency graph whose manager dispatches parallel subagents and revises tasks as new information arrives. Desktop and web experiments examine coordination, information handoff and task-completion time.", "topics": ["planning", "long-horizon", "workflow", "state-tracking", "efficiency"], "updated_at": "2026-06-01", "url": "https://arxiv.org/abs/2606.01533"}
{"aliases": [], "authors": ["Yuhang Han", "Wenzheng Yang", "Yujie Chen", "Xiangqi Jin", "Yaojie Zhang", "Siteng Huang", "Linfeng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73364572c0f21976048c89cd7732a16c7a952b08986792d02681a5ad0288c3ca", "submitted_at": "2026-06-01T07:10:00Z", "updated_at": "2026-06-01T07:10:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.01790v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2606.01790", "identifiers": {"arxiv": "2606.01790"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.01790"}], "name": "STaR-KV: Spatio-Temporal Adaptive Re-weighting for KV Cache Compression in GUI Vision-Language Models", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "STaR-KV compresses GUI-agent KV caches by calibrating visual-token importance across attention subspaces and interaction time. Its training-free scoring combines spatial specialization, temporal redundancy and an adaptive temperature, evaluated at matched cache budgets.", "topics": ["efficiency", "training-free", "screen-understanding", "interaction-history"], "updated_at": "2026-06-01", "url": "https://arxiv.org/abs/2606.01790"}
{"aliases": [], "authors": ["Junqi Liu", "Selena Song", "Yuhan Wang", "Jiawei Mao", "Hardy Chen", "Xiaoke Huang", "Tianhao Qi", "Pengfei Guo", "Yucheng Tang", "Yufan He", "Can Zhao", "Andriy Myronenko", "Dong Yang", "Daguang Xu", "Yuyin Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a3f2718605795e8a8b617ae3bf23c32a7f265d584f1c7c051c58030b38ba13e", "submitted_at": "2026-06-01T09:22:55Z", "updated_at": "2026-06-03T05:43:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.01961v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.01961", "identifiers": {"arxiv": "2606.01961"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.01961"}], "name": "AutoMedBench: Towards Medical AutoResearch with Agentic AI Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-03", "url": "https://arxiv.org/abs/2606.01961"}
{"aliases": [], "authors": ["Rui Yang", "Qianhui Wu", "Yuxi Chen", "Hao Bai", "Wenlin Yao", "Hao Cheng", "Baolin Peng", "Huan Zhang", "Tong Zhang", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4d5fdcffb29f45407f41a5e00d7c4052051b05f4ca5b320feffad9e8f070b7c2", "submitted_at": "2026-06-01T10:20:10Z", "updated_at": "2026-06-04T09:03:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.02031v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.02031", "identifiers": {"arxiv": "2606.02031"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.02031"}], "name": "OpenWebRL: Demystifying Online Multi-turn Reinforcement Learning for Visual Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2606.02031"}
{"aliases": [], "authors": ["Athar Parvez", "Muhammad Jawad Mufti", "Muqaddas Gull", "Omar Hammad"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0b3bc81cb9da22783c39a05f459d242115c05d28c433e88f6e3df74c1a49caf0", "submitted_at": "2026-06-01T13:08:07Z", "updated_at": "2026-06-01T13:08:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.02208v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2606.02208", "identifiers": {"arxiv": "2606.02208"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.02208"}], "name": "Context-Aware Workflow Decomposition for Automated Mobile UI Annotation Using Multimodal Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-01", "url": "https://arxiv.org/abs/2606.02208"}
{"aliases": [], "authors": ["Yuting Ning", "Zhehao Zhang", "Yash Kumar Lal", "Boyu Gou", "Junyi Li", "Weitong Ruan", "Chentao Ye", "Rahul Gupta", "Diyi Yang", "Yu Su", "Huan Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "731e5f8e51ca039b8c06a37571b70d9581da37ce7a097a34eb8fcc722f80ad7b", "submitted_at": "2026-06-01T17:45:39Z", "updated_at": "2026-10-11T22:23:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.02540v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.02540", "identifiers": {"arxiv": "2606.02540"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.02540"}], "name": "SkillHarm: Lifecycle-Aware Skill-Based Attacks via Automated Construction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-11", "url": "https://arxiv.org/abs/2606.02540"}
{"aliases": [], "authors": ["Wenkai Wang", "Tao Xiong", "Jingchen Ni", "Yunpeng Bao", "Xiyun Li", "Tianqi Liu", "Hongcan Guo", "Zilong Huang", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf0fcafeb2d40e489ec80784a3541114b02c6a9077a330265d5c49775eca4582", "submitted_at": "2026-06-02T03:42:34Z", "updated_at": "2026-06-02T03:42:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.03103v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.03103", "identifiers": {"arxiv": "2606.03103"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.03103"}], "name": "DeskCraft: Benchmarking Desktop Agents on Professional Workflows and Human-in-the-Loop Collaboration", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-06-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "DeskCraft evaluates professional desktop workflows and human-agent collaboration through mid-execution clarification, interruption and post-completion feedback. Its tasks cover creative applications and expose difficulties in long-horizon execution and proactive clarification.", "topics": ["evaluation", "human-in-the-loop", "long-horizon", "workflow"], "updated_at": "2026-06-02", "url": "https://arxiv.org/abs/2606.03103"}
{"aliases": [], "authors": ["Jia Yu", "Zilong Wang", "Xinyang Jiang", "Dongsheng Li", "Shuo Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b23c12cafdd2d4a3ef63027c914ed3894c61c555edd89468c56ae40c3ec6cea9", "submitted_at": "2026-06-02T06:02:35Z", "updated_at": "2026-06-02T06:02:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.03203v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.03203", "identifiers": {"arxiv": "2606.03203"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.03203"}], "name": "MedCUA-Bench: A Screenshot-Only Benchmark for Clinical Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "MedCUA-Bench evaluates screenshot-only clinical interface workflows using paired intent- and step-level goals. It combines reconstructed medical interfaces with completion and clinical-safety checks, and separately reports results on a real OpenEMR system.", "topics": ["evaluation", "safety-security", "science", "workflow"], "updated_at": "2026-06-02", "url": "https://arxiv.org/abs/2606.03203"}
{"aliases": [], "authors": ["Zhijie Ding", "Weinan Hong", "Zicheng Zhu", "Lei Li", "Dezhi Kong", "Hao Wang", "Peng Zhou", "Xuchu Jiang", "Jiaming Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df42655cf7b061c55e50014ac1b66fb6580e9f54aac1efd0aa6c7f108b900a26", "submitted_at": "2026-06-02T06:54:02Z", "updated_at": "2026-08-31T12:02:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.03236v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2606.03236", "identifiers": {"arxiv": "2606.03236"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.03236"}], "name": "Perceive Before Reasoning: A Pre-Reasoning Perception Framework for Efficient and Reliable Proactive Mobile Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2606.03236"}
{"aliases": [], "authors": ["Yuhao Yang", "Tianyu Fan", "Chao Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f7490150e82a5abfc510c60401cccf21c821d829e66cc5f2b40fd2fbbf4b0df4", "submitted_at": "2026-06-02T16:30:33Z", "updated_at": "2026-06-02T16:30:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.03854v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.03854", "identifiers": {"arxiv": "2606.03854"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.03854"}], "name": "CLI-Anything: Towards Agent-Native Computer Use", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced. Adjacent computer-interface architecture: the described application command harnesses concern computer operation, but no GUI execution or comparative benchmark result is established in the abstract.", "published_at": "2026-06-02", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "CLI-Anything argues for application interfaces built around structured commands, explicit state and execution feedback. It describes transforming existing software into command-line harnesses and organizing that approach through CLI-Hub.", "topics": ["automation", "tool-integration", "workflow"], "updated_at": "2026-06-02", "url": "https://arxiv.org/abs/2606.03854"}
{"aliases": [], "authors": ["Zongwei Lv", "Yaoming Li", "Zhewen Tan", "Yilun Yao", "Yuxuan Tian", "Lin Sun", "Xiangzheng Zhang", "Weihong Lin", "Tong Yang", "Guangxiang Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "96504157ef371a21ed496b5f73fcf8358137b069fcc76405c81b1a6fb3e97591", "submitted_at": "2026-06-02T16:51:24Z", "updated_at": "2026-08-28T05:22:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.03889v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.03889", "identifiers": {"arxiv": "2606.03889"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.03889"}], "name": "RealClawBench: Live OpenClaw Benchmarks from Real Developer-Agent Sessions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-28", "url": "https://arxiv.org/abs/2606.03889"}
{"aliases": [], "authors": ["Yingqi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc55e544d43ae80db777d5ccc1b151ecadc6bc61a1fca735390529f68251fbff", "submitted_at": "2026-06-02T16:53:24Z", "updated_at": "2026-08-18T06:50:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.03895v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.03895", "identifiers": {"arxiv": "2606.03895"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.03895"}], "name": "Agent libOS: A Runtime Substrate for Capability-Controlled Self-Evolving LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-18", "url": "https://arxiv.org/abs/2606.03895"}
{"aliases": [], "authors": ["Zechen Bai", "Zhiheng Chen", "Yiqi Lin", "Kevin Qinghong Lin", "Difei Gao", "Xiangwu Guo", "Xin Wang", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "20baed183b31118c23112d6396ad9ea15329ee8b54ef435caf003907f18d9c28", "submitted_at": "2026-06-02T17:39:07Z", "updated_at": "2026-06-02T17:39:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.03951v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2606.03951", "identifiers": {"arxiv": "2606.03951"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.03951"}], "name": "Demo2Tutorial: From Human Experience to Multimodal Software Tutorials", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-02", "url": "https://arxiv.org/abs/2606.03951"}
{"aliases": [], "authors": ["Zhikai Chen", "Jialiang Gu", "Junyu Yin", "Xianxuan Long", "Shenglai Zeng", "Xiaoze Liu", "Kai Guo", "Keren Zhou", "Jiliang Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "72d2f12a093ccda61853be0a4c20d19eed1203518bd1c544f32417c7ec1c33ed", "submitted_at": "2026-06-03T00:42:49Z", "updated_at": "2026-06-03T00:42:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.04315v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.04315", "identifiers": {"arxiv": "2606.04315"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.04315"}], "name": "Exploring Cross-Scenario Generality of Agentic Memory Systems: Diagnostics and a Strong Baseline", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-03", "url": "https://arxiv.org/abs/2606.04315"}
{"aliases": [], "authors": ["Jiaxi Li", "Ke Deng", "Yun Wang", "Jingyuan Huang", "Yucheng Shi", "Qiaoyu Tan", "Jin Lu", "Ninghao Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d891956bde9b8672859fcfcd2d54dfd3e4720964c25d77cc28bc1200ef3a56c", "submitted_at": "2026-06-03T03:11:50Z", "updated_at": "2026-06-03T03:11:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.04391v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.04391", "identifiers": {"arxiv": "2606.04391"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.04391"}], "name": "Online Skill Learning for Web Agents via State-Grounded Dynamic Retrieval", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "State-Grounded Dynamic Retrieval extracts reusable sub-procedures from completed web trajectories and represents them as paired text and code. Skills are retrieved against both the current webpage state and the task goal at intermediate execution steps.", "topics": ["web-navigation", "memory", "interaction-history", "planning", "workflow"], "updated_at": "2026-06-03", "url": "https://arxiv.org/abs/2606.04391"}
{"aliases": [], "authors": ["Xinyu Lu", "Tianshu Wang", "Pengbo Wang", "zujie wen", "Zhiqiang Zhang", "Jun Zhou", "Boxi Cao", "Yaojie Lu", "Hongyu Lin", "Xianpei Han", "Le Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce66de481b753963151690a09f5ff10800e2ee671afe05bb3b68140f30947759", "submitted_at": "2026-06-03T04:58:17Z", "updated_at": "2026-06-03T04:58:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.04455v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.04455", "identifiers": {"arxiv": "2606.04455"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.04455"}], "name": "The Meta-Agent Challenge: Are Current Agents Capable of Autonomous Agent Development?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-03", "url": "https://arxiv.org/abs/2606.04455"}
{"aliases": [], "authors": ["Zhichao Yang", "Yuanze Hu", "Haojie Hao", "Longkun Hao", "Dongshuo Huang", "Hongyu Lin", "Gen Li", "Lanqing Hong", "Yihang Lou", "Yan Bai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dcb2d643f4c802180577ec03ac8a3f42f8561d98a7008a0abd53c4f9b6f2348f", "submitted_at": "2026-06-03T09:01:24Z", "updated_at": "2026-06-06T08:34:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.04627v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2606.04627", "identifiers": {"arxiv": "2606.04627"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.04627"}], "name": "MIRAGE: Mobile Agents with Implicit Reasoning and Generative World Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-06", "url": "https://arxiv.org/abs/2606.04627"}
{"aliases": [], "authors": ["Jiashu Yao", "Heyan Huang", "Daiqing Wu", "Wangke Chen", "Huaxi Ai", "Haoyu Wen", "Zeming Liu", "Yuhang Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4d5a90333c4f6be0b8d203a2f5938f7e7eefcb09fbe7ea47b85eba8c7c1c422a", "submitted_at": "2026-06-03T10:25:46Z", "updated_at": "2026-06-03T10:25:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.04701v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.04701", "identifiers": {"arxiv": "2606.04701"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.04701"}], "name": "Benchmarking Living-Screen-Native GUI Agents on Short-Video Platforms", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "LivingScreen evaluates agents in a browser-based short-video environment where observing content consumes time and affects subsequent actions. Its task suite scores accuracy together with information efficiency to diagnose over- and under-observation on continuously changing screens.", "topics": ["evaluation", "video", "efficiency", "screen-understanding"], "updated_at": "2026-06-03", "url": "https://arxiv.org/abs/2606.04701"}
{"aliases": [], "authors": ["Nicholas Saban"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b649c0cef63bc15558aa4498a65043644c109faaf6403b610c2132fd3f5706d", "submitted_at": "2026-06-03T01:21:59Z", "updated_at": "2026-06-03T01:21:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.05233v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.05233", "identifiers": {"arxiv": "2606.05233"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.05233"}], "name": "Domain-Conditioned Safety in Frontier Computer-Using Agents: A 793-Episode Browser Benchmark, a Coding-Domain Cross-Reference, and a Reproducibility Audit of Recent Red-Teaming", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CUA-HandCrafted, a browser-agent red-teaming benchmark that varies attack templates, tasks and system prompts. Its domain comparison and reproducibility audit caution against transferring browser-surface safety conclusions to coding agents or unreleased optimized injections.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-06-03", "url": "https://arxiv.org/abs/2606.05233"}
{"aliases": [], "authors": ["Kuangshi Ai", "Haichao Miao", "Kaiyuan Tang", "Shusen Liu", "Chaoli Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40169820467c7cf29f1b30068722907ad5a9fa482002645afd99dadf0808d038", "submitted_at": "2026-06-04T00:14:25Z", "updated_at": "2026-06-04T00:14:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.05525v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "id": "arxiv:2606.05525", "identifiers": {"arxiv": "2606.05525"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.05525"}], "name": "SciVisAgentSkills: Design and Evaluation of Agent Skills for Scientific Data Analysis and Visualization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2606.05525"}
{"aliases": [], "authors": ["Hao Bai", "Rui Yang", "Chenlu Ye", "Spencer Whitehead", "Aviral Kumar", "Tong Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a8afe44ffa73739505070c5f8004d8b4bbe98455fc0f5f3cef4b450bf441c74", "submitted_at": "2026-06-04T02:18:44Z", "updated_at": "2026-08-07T17:58:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.05597v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.05597", "identifiers": {"arxiv": "2606.05597"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.05597"}], "name": "AsyncWebRL: Efficient Asynchronous Reinforcement Learning for Multi-Step Visual Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "AsyncWebRL overlaps web-agent rollouts, updates and policy refresh while managing screenshot handling and a persistent rollout pool. It also replaces trajectory-length normalization in multi-step GRPO to reduce verbose failed trajectories, with throughput and out-of-distribution evaluation.", "topics": ["reinforcement-learning", "efficiency", "web-navigation"], "updated_at": "2026-08-07", "url": "https://arxiv.org/abs/2606.05597"}
{"aliases": [], "authors": ["Xuehang Guo", "Zora Zhiruo Wang", "Qingyun Wang", "Graham Neubig", "Xingyao Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "323a776c9cce297d75b8ec0550fa4803fe370186195e96b265ca24949af331b1", "submitted_at": "2026-06-04T03:17:21Z", "updated_at": "2026-06-04T03:17:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.05646v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.05646", "identifiers": {"arxiv": "2606.05646"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.05646"}], "name": "Enhancing Software Engineering Through Closed-Loop Memory Optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2606.05646"}
{"aliases": [], "authors": ["Yuhang Fu", "Ruishan Fang", "Jiaqi Shao", "Huiyu Zheng", "Zhengtao Zhu", "Bing Luo", "Zhenyu Wen", "Tao Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a016c6c65375136ae12b5dd49d1b72eb10a0fb7306df82fa82428c53a2db7fb9", "submitted_at": "2026-06-04T03:50:47Z", "updated_at": "2026-10-09T07:41:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.05670v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.05670", "identifiers": {"arxiv": "2606.05670"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.05670"}], "name": "Do More Agents Help? Evaluating Accuracy and Cost in LLM Multi-Agent Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2606.05670"}
{"aliases": [], "authors": ["Nicolas Bougie", "Xiaotong Ye", "Gian Maria Marconi", "Narimasa Watanabe"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1aae315d74a497af3c4862b1d8712d49c5ec8b3fb9302e3ea0eaf6692d85e0a7", "submitted_at": "2026-06-04T04:35:16Z", "updated_at": "2026-06-04T04:35:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.05697v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2606.05697", "identifiers": {"arxiv": "2606.05697"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.05697"}], "name": "PerceptUI: LLM Agents as Human-Aligned Synthetic Users for UI/UX Evaluation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2606.05697"}
{"aliases": [], "authors": ["Nathan Bout", "Maxime Langevin", "Ronan Riochet"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27ffcb10d966946de9fa2353c4b96798e5f090bdfdc5610f5e8a5c6233cf70ea", "submitted_at": "2026-06-04T15:57:29Z", "updated_at": "2026-06-04T15:57:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.06322v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}], "id": "arxiv:2606.06322", "identifiers": {"arxiv": "2606.06322"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.06322"}], "name": "DragOn: A Benchmark and Dataset for Drag-Based GUI Interactions", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "DragOn supplies drag-grounding training data and a held-out benchmark for highlighting text, selecting cells, resizing elements and moving sliders. It evaluates existing vision-language models and a model fine-tuned on the drag tasks.", "topics": ["grounding", "data-generation", "evaluation", "screen-understanding"], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2606.06322"}
{"aliases": [], "authors": ["Yasmine Omri", "Ziyu Gan", "Zachary Broveak", "Robin Geens", "Zexue He", "Alex Pentland", "Marian Verhelst", "Tsachy Weissman", "Thierry Tambe"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b97e46f1c9ef1f5ecac50cc0a0a918860dd1e76f8016ba64c459db7ea095186", "submitted_at": "2026-06-04T17:44:18Z", "updated_at": "2026-09-22T17:40:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.06448v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2606.06448", "identifiers": {"arxiv": "2606.06448"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.06448"}], "name": "Agent Memory: Characterization and System Implications of Stateful Long-Horizon Workloads", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-22", "url": "https://arxiv.org/abs/2606.06448"}
{"aliases": [], "authors": ["Shiyun Xiong", "Dongming Wu", "Peiwen Sun", "Yuang Ai", "Bokang Yang", "Wencheng Han", "Xiao-Hui Li", "Xiangyu Yue"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c67a0aaaa5319694d0aeb6bf1715c0ef2d349f38c1f5b9a79cd347585c326bc6", "submitted_at": "2026-06-04T17:52:04Z", "updated_at": "2026-06-04T17:52:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.06462v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2606.06462", "identifiers": {"arxiv": "2606.06462"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.06462"}], "name": "Benchmark Everything Everywhere All at Once", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2606.06462"}
{"aliases": [], "authors": ["Victor Muryn", "Maksym Shamrai", "Sofiia Mazepa", "Yehor Khodysko"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd66adb4e99df2348f59e238d219d48cb6e758e694839367d2b9a64675429d87", "submitted_at": "2026-06-04T14:01:32Z", "updated_at": "2026-06-04T14:01:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.06560v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.06560", "identifiers": {"arxiv": "2606.06560"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.06560"}], "name": "MacArena: Benchmarking Computer Use Agents on an Online macOS Environment", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "MacArena runs computer-use tasks through Apple’s virtualization framework on Apple Silicon, combining task ports with macOS-native workloads. It distinguishes the ported and native subsets when examining agent performance and cross-platform transfer.", "topics": ["evaluation", "robustness"], "updated_at": "2026-06-04", "url": "https://arxiv.org/abs/2606.06560"}
{"aliases": [], "authors": ["Shubham Gaur", "Ian Lane"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a449bb2706d5d18cbd176dcfd2f14d814cf1c1a22a08e9210c99e0722e10ea93", "submitted_at": "2026-06-04T20:48:37Z", "updated_at": "2026-08-04T18:57:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.06708v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.06708", "identifiers": {"arxiv": "2606.06708"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.06708"}], "name": "Signal-Driven Observation for Long-Horizon Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced. The abstract presents an architecture proposal and open problems, without reporting an empirical evaluation.", "published_at": "2026-06-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes Signal-Driven Observation: a separate call selects relevant DOM elements and selectors, and a detector requests new observations only after meaningful browser events or failures. The paper frames observation timing and compression as an architectural research problem for long-horizon web agents.", "topics": ["efficiency", "screen-understanding", "long-horizon"], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2606.06708"}
{"aliases": [], "authors": ["Hanxun Yu", "Xuan Qu", "Lei Ke", "Boqiang Zhang", "Yuxin Wang", "Jianke Zhu", "Dong Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "efb60a681d20749124c2fbff1c21161dbde65f29913fd9efa3de553f816b5b0d", "submitted_at": "2026-06-05T04:16:24Z", "updated_at": "2026-06-05T04:16:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.06891v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.06891", "identifiers": {"arxiv": "2606.06891"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.06891"}], "name": "Stream3D-VLM: Online 3D Spatial Understanding with Incremental Geometry Priors", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-05", "url": "https://arxiv.org/abs/2606.06891"}
{"aliases": [], "authors": ["Haojie Hao", "Longkun Hao", "Yihang Lou", "Yan Bai", "Zhenyang Li", "Zhichao Yang", "Dongshuo Huang", "Hongyu Lin", "Lanqing Hong", "Jiakai Wang", "Xianglong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d530152b210a69cb458fecb036c10621c72daccfbcac7c8306c7e248c87cf14", "submitted_at": "2026-06-05T08:17:28Z", "updated_at": "2026-06-12T07:28:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.07027v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.07027", "identifiers": {"arxiv": "2606.07027"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.07027"}], "name": "StainFlow: Entity-Stain Tracking and Evidence Linking for Process Rewards in GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "StainFlow tracks visually verifiable task entities through GUI trajectories to identify phases and retrieve focused evidence windows. Its process-reward model is tested for both online policy learning and trajectory-completion judgments.", "topics": ["reward-models", "state-tracking", "reinforcement-learning", "evaluation"], "updated_at": "2026-06-12", "url": "https://arxiv.org/abs/2606.07027"}
{"aliases": [], "authors": ["Zequn Xie", "Junjie Wang", "Dan Yang", "Jie Feng", "Yue Shen", "Jian Wang", "Jinjie Gu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2dfd022afb88507a33786727f47e2a3b1bee590a7558de6b8ce3e6f8c14a46cb", "submitted_at": "2026-06-05T09:10:50Z", "updated_at": "2026-06-05T09:10:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.07074v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.07074", "identifiers": {"arxiv": "2606.07074"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.07074"}], "name": "SlimSearcher: Training Efficiency-Aware Web Agents via Adaptive Reward Gating", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-05", "url": "https://arxiv.org/abs/2606.07074"}
{"aliases": [], "authors": ["Yudi Zhang", "Meng Fang", "Zhenfang Chen", "Mykola Pechenizkiy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b94b2c62b2206d95323fb1a21d85e41ffcafdd4a67a5d7aa704853efd157ddc2", "submitted_at": "2026-06-05T15:09:52Z", "updated_at": "2026-06-05T15:09:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.07367v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.07367", "identifiers": {"arxiv": "2606.07367"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.07367"}], "name": "Self-evolving LLM agents with in-distribution Optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-05", "url": "https://arxiv.org/abs/2606.07367"}
{"aliases": [], "authors": ["Bo Zhang", "Borui Zhang", "Chenghao Jiang", "Minglei Shi", "Xiaofeng Wang", "Zheng Zhu", "Jie Zhou", "Jiwen Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33c343e9bc02976fc3e32057cdcdebc3abeae5ef049b877443e81917a8abc550", "submitted_at": "2026-05-28T17:59:31Z", "updated_at": "2026-05-28T17:59:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.07594v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.07594", "identifiers": {"arxiv": "2606.07594"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.07594"}], "name": "Syll: Open-Source Personal Automation with Cross-Surface Execution", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-05-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Syll combines tool calls, shell execution and visual GUI control in a personal-automation runtime. It turns user demonstrations into reusable skills and exposes execution logs, keyframes and approval checkpoints for inspection.", "topics": ["automation", "human-in-the-loop", "workflow", "memory", "recording", "observability"], "updated_at": "2026-05-28", "url": "https://arxiv.org/abs/2606.07594"}
{"aliases": [], "authors": ["Rishi Desai", "Jesse Hu", "Joan Cabezas", "Neel Harsola", "Pratyush Shukla", "Roey Ben Chaim", "Adnan El Assadi", "Omkaar Mukund Kamath", "Fenil Faldu", "Prannay Hebbar", "Jiankai Sun", "Yiyuan Li", "Pramod Srinivasan", "Ishan Gupta", "Christopher Settles", "Daniel Wang", "Derek Chen", "Pranav Raja", "Albert Liu", "Marek Šuppa", "Nevasini Sasikumar", "Luyang Kong", "Erik Quintanilla", "Xiangyi Li", "Ivan Bercovich", "Steven Dillmann"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "00555477a43df2aba6d24761bcefd77f3499299faf97a146d65aaf634476b131", "submitted_at": "2026-06-05T00:39:44Z", "updated_at": "2026-06-05T00:39:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.07682v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.07682", "identifiers": {"arxiv": "2606.07682"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.07682"}], "name": "SWE-Marathon: Can Agents Autonomously Complete Ultra-Long-Horizon Software Work?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-05", "url": "https://arxiv.org/abs/2606.07682"}
{"aliases": [], "authors": ["Chung-En Sun", "Linbo Liu", "Tsui-Wei Weng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f7fd6a97902088c81d79df455718f8788361c16ac1030b363b6988c98dc664a2", "submitted_at": "2026-06-05T21:58:19Z", "updated_at": "2026-08-27T04:42:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.07867v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.07867", "identifiers": {"arxiv": "2606.07867"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.07867"}], "name": "The Cold-Start Safety Gap in LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-27", "url": "https://arxiv.org/abs/2606.07867"}
{"aliases": [], "authors": ["Radeen Mostafa", "Sawradip Saha"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d29dec745923e652654521f65a32f919a99d4577992006efc1a5e469bf775d81", "submitted_at": "2026-06-08T12:18:22Z", "updated_at": "2026-06-08T12:18:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.09399v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2606.09399", "identifiers": {"arxiv": "2606.09399"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.09399"}], "name": "RunAgent SuperBrowser: A Theory of Autonomous Web Navigation Grounded in Human Browsing Behaviour", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "SUPERBROWSER combines screenshot-based target proposals with separate orchestration, planning and action roles. A bounded ledger and context-eviction process retain recent actions, task facts and checkpoints during browser execution.", "topics": ["web-navigation", "grounding", "planning", "memory", "efficiency"], "updated_at": "2026-06-08", "url": "https://arxiv.org/abs/2606.09399"}
{"aliases": [], "authors": ["Wanli Li", "Bowen Zhou", "Yunyao Yu", "Zhou Xu", "Yifan Yang", "Dongsheng Li", "Caihua Shan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73b751efd63d98fc81166a1a80cb0c162d68bf1ec4ecca0a1a1a7717ad62e680", "submitted_at": "2026-06-08T12:39:23Z", "updated_at": "2026-07-06T14:25:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.09426v3"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.09426", "identifiers": {"arxiv": "2606.09426"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.09426"}], "name": "WeaveBench: A Long-Horizon, Real-World Benchmark for Computer-Use Agents with Hybrid Interfaces", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "WeaveBench requires agents to combine GUI, command-line and code operations within long tasks on an Ubuntu desktop. A trajectory-aware judge checks deliverables and interaction evidence, including shortcut behavior that outcome-only grading can miss.", "topics": ["evaluation", "long-horizon", "workflow", "tool-integration"], "updated_at": "2026-07-06", "url": "https://arxiv.org/abs/2606.09426"}
{"aliases": [], "authors": ["Bojie Rong", "Zheyu Shen", "Qiaoping Wang", "Pengfei Kang", "Yang Xu", "Yawen Wei", "Hanyu Wu", "Zhi Zhao", "Leihao Pei", "Linquan Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a852c77650fa54f234794c6cd27777f4e7219b8389b509d807e27db16138a1b", "submitted_at": "2026-06-08T12:55:42Z", "updated_at": "2026-06-08T12:55:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.09447v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.09447", "identifiers": {"arxiv": "2606.09447"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.09447"}], "name": "AliyunConsoleAgent: Training Web Agents in Real-World Cloud Environments via Distillation and Reinforcement Learning", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "AliyunConsoleAgent trains browser agents to verify documented procedures against changing cloud-console UIs through trajectory distillation and online GRPO. Provisioned rollout environments and backend audit-log checks support training and outcome evaluation.", "topics": ["web-navigation", "imitation-learning", "reinforcement-learning", "evaluation"], "updated_at": "2026-06-08", "url": "https://arxiv.org/abs/2606.09447"}
{"aliases": [], "authors": ["Lawrence Keunho Jang", "Mareks Woodside", "Geronimo Carom", "Andrew Keunwoo Jang", "Jing Yu Koh", "Ruslan Salakhutdinov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "495420bd87ba06ae2ec1763865b1145271e964b4bd40145c8e591653a1581c42", "submitted_at": "2026-06-08T17:27:13Z", "updated_at": "2026-06-08T17:27:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.09764v1"}}, "environments": ["ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "id": "arxiv:2606.09764", "identifiers": {"arxiv": "2606.09764"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.09764"}], "name": "iOSWorld: A Benchmark for Personally Intelligent Phone Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces a native iOS simulator benchmark with persistent user data, multi-app tasks, and personalization-dependent phone workflows.", "topics": ["evaluation", "cross-app", "memory"], "updated_at": "2026-06-08", "url": "https://arxiv.org/abs/2606.09764"}
{"aliases": [], "authors": ["Sirui Liang", "Bohan Yu", "Peiyu Wang", "Shiguang Guo", "Wenxing Hu", "Pengfei Cao", "Jian Zhao", "Cao Liu", "Ke Zeng", "Xunliang Cai", "Kang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "884791f403d3f46da6dd458939c60aa4072020beeb06ed4548f67fdbcd9789f3", "submitted_at": "2026-06-09T04:16:35Z", "updated_at": "2026-06-09T04:16:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.10394v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.10394", "identifiers": {"arxiv": "2606.10394"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.10394"}], "name": "STAGE-Claw: Automated State-based Agent Benchmarking for Realistic Scenarios", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-09", "url": "https://arxiv.org/abs/2606.10394"}
{"aliases": [], "authors": ["Jayoo Hwang", "Xiaowen Zhang", "Vedant Padwal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "386bd82be97d5cbe64fdb3478e427362e710d78d423c9e52661a8a96ea2fdecf", "submitted_at": "2026-06-09T04:53:19Z", "updated_at": "2026-06-09T04:53:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.10423v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.10423", "identifiers": {"arxiv": "2606.10423"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.10423"}], "name": "WebChallenger: A Reliable and Efficient Generalist Web Agent", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebChallenger builds a hierarchical DOM-derived page representation for selective observation, reusable site maps and compound actions. The framework combines these mechanisms without model fine-tuning and evaluates generalization across multiple web-agent benchmarks.", "topics": ["web-navigation", "memory", "efficiency", "training-free", "workflow"], "updated_at": "2026-06-09", "url": "https://arxiv.org/abs/2606.10423"}
{"aliases": [], "authors": ["Can Lin", "Tao Feng", "Hangjie Yuan", "Dan Zhang", "Yifan Zhu", "Zhonghong Ou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7523d3901755d8feeecbaf2f02dcc00fd0a40102e1498d49b9f789b0a7c02d75", "submitted_at": "2026-06-09T07:52:10Z", "updated_at": "2026-07-06T06:23:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.10522v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.10522", "identifiers": {"arxiv": "2606.10522"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.10522"}], "name": "GUI-AC: Enhancing Continual Learning in GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-AC uses grounding certainty to adjust advantage weighting and policy-clipping bounds during continual reinforcement fine-tuning. Its mechanisms aim to suppress noisy confidence updates while retaining exploration under shifting GUI domains and resolutions.", "topics": ["reinforcement-learning", "grounding", "robustness"], "updated_at": "2026-07-06", "url": "https://arxiv.org/abs/2606.10522"}
{"aliases": [], "authors": ["David Hofer", "Edoardo Debenedetti", "Florian Tramèr"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "195339379de38edfb17976df437115b4a443ee821d4645d78a06dfaccee7c8f6", "submitted_at": "2026-06-09T07:54:58Z", "updated_at": "2026-06-09T07:54:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.10525v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.10525", "identifiers": {"arxiv": "2606.10525"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.10525"}], "name": "Assessing Automated Prompt Injection Attacks in Agentic Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-09", "url": "https://arxiv.org/abs/2606.10525"}
{"aliases": [], "authors": ["Yv Zhang", "Hao Sun", "Hao Fang", "Kuofeng Gao", "Fan Mo", "Bin Chen", "Shu-Tao Xia", "Yaowei Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "16c0eecd3208d9d4bc2ce806682e73ea6baf4691404595d2d57476a6a42facad", "submitted_at": "2026-06-09T11:53:25Z", "updated_at": "2026-06-09T11:53:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.10742v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.10742", "identifiers": {"arxiv": "2606.10742"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.10742"}], "name": "MemVenom: Triggered Poisoning of Multimodal Memories in Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-09", "url": "https://arxiv.org/abs/2606.10742"}
{"aliases": [], "authors": ["Liya Zhu", "Jingzhe Ding", "Jian Zhang", "Jianbo Xue", "Shihao Liang", "Ge Zhang", "Yi Zhu", "Duju Zeng", "Xiang Gao", "Qingshui Gu", "Mailun Gao", "Huimin Che", "Yan Zhao", "Peiheng Zhou", "Haojun Wang", "Chaobo Xian", "Lili Le", "Chi Wu", "Yiwei Liu", "Shengda Long", "Jiale Yang", "Fangzhi Xu", "Sijin Wu", "Haodong Duan", "Chao He", "Zhaojian Li", "Minchao Wang", "Huan Zhou", "Jiani Hou", "Chuqian Yu", "Weiran Shi", "Hongwan Gao", "Jiamin Chen", "Guanhong Chen", "Tingqin Luo", "Kaiyuan Zhang", "Zhixin Yao", "Qing Hua", "Yuhao Jiang", "Jin Chen", "Pu Chen", "Zhenyu Hu", "Xingyu Li", "Zhengxuan Jiang", "Meng Cao", "Tianfeng Long", "Haozhe Wang", "Mingzhang Wang", "Yichen Zhang", "Yiming Dai", "Chenchen Zhang", "Jiaying Wang", "Xinying Liu", "Xingzu Liu", "Lingling Zhang", "Xinjie Chen", "Yujia Qin", "Wangchunshu Zhou", "Zhiyong Wu", "Yang Liu", "Jiaheng Liu", "Lei Zhang", "Shen Yan", "Wenhao Huang", "Zaiyuan Wang", "Xiaolong Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "037a3197137d27318320a3fd23fb5e606a1aa748133e57213cd02fd10aad4c4d", "submitted_at": "2026-06-09T16:10:16Z", "updated_at": "2026-07-17T12:22:12Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2606.11042v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.11042", "identifiers": {"arxiv": "2606.11042"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.11042"}], "name": "Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Workflow-GYM evaluates long GUI workflows in professional software rather than isolated interface actions. Its analysis examines failures such as omitted stages, accumulated errors and goal drift during specialized end-to-end work.", "topics": ["evaluation", "long-horizon", "workflow", "robustness"], "updated_at": "2026-07-17", "url": "https://arxiv.org/abs/2606.11042"}
{"aliases": [], "authors": ["Jaewoo Lee", "Zaid Khan", "Archiki Prasad", "Justin Chih-Yao Chen", "Supriyo Chakraborty", "Kartik Balasubramaniam", "Sambit Sahu", "Elias Stengel-Eskin", "Hyunji Lee", "Mohit Bansal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8ed399c8b489a80c5313a00232d2ecca0fbc8228b1f56745c02d4fedb6f6908", "submitted_at": "2026-06-09T16:39:10Z", "updated_at": "2026-06-09T16:39:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.11078v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.11078", "identifiers": {"arxiv": "2606.11078"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.11078"}], "name": "A History-Aware Visually Grounded Critic for Computer Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "HiViG combines a compact history of completed actions with a visually grounded critic that checks proposed coordinates before execution. Its evaluations separate history-related planning failures from errors caused by targeting the wrong screen element.", "topics": ["interaction-history", "memory", "grounding", "reward-models", "long-horizon", "evaluation"], "updated_at": "2026-06-09", "url": "https://arxiv.org/abs/2606.11078"}
{"aliases": [], "authors": ["Lingzhi Yuan", "Chenghao Deng", "Fangxu Yu", "Souradip Chakraborty", "Mohammad Rostami", "Furong Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c6ae7fba27ac56615f082e6bc211e55654f0ff50e504f5c5e55184bbe6455d38", "submitted_at": "2026-06-09T17:58:21Z", "updated_at": "2026-06-09T17:58:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.11290v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.11290", "identifiers": {"arxiv": "2606.11290"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.11290"}], "name": "FlowBank: Query-Adaptive Agentic Workflows Optimization through Precompute-and-Reuse", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-09", "url": "https://arxiv.org/abs/2606.11290"}
{"aliases": [], "authors": ["Siyuan Luo", "Nairong Zheng", "Lin Zhou", "Tiankuo Yao", "Shengyou Yuan", "Haojia Yu", "Cong Pang", "Jiapeng Luo", "Lewei Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ddaaf4319fde0390c0caa19f5f9acd42049fa8c105f5f2d85594ed123548cdd3", "submitted_at": "2026-06-09T23:44:26Z", "updated_at": "2026-07-14T14:04:47Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2606.11520v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.11520", "identifiers": {"arxiv": "2606.11520"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.11520"}], "name": "ISE: An Execution-Grounded Recipe for Multi-Turn OS-Agent Trajectories", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-14", "url": "https://arxiv.org/abs/2606.11520"}
{"aliases": [], "authors": ["Changyue Wang", "Weihang Su", "Qingyao Ai", "Yichen Tang", "Runzhong Qiao", "Xuancheng Li", "Min Zhang", "Yiqun Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "70f57e75cc7a03863906f4cf43d44b44c85e5ce2477da948bb345a3ba0532507", "submitted_at": "2026-06-10T15:21:18Z", "updated_at": "2026-06-10T15:21:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.12203v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.12203", "identifiers": {"arxiv": "2606.12203"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.12203"}], "name": "Adaptive Multi-Resolution Procedural Knowledge Compression for Large Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-10", "url": "https://arxiv.org/abs/2606.12203"}
{"aliases": [], "authors": ["Longkun Hao", "Hongyu Lin", "Hao Li", "Zhuowen Liu", "Zhichao Yang", "Haojie Hao", "Dongshuo Huang", "Haitao Yang", "Hongyu Ge", "Ming jie Xie", "Yanjun Wu", "Zi Hao Yin", "Yan Bai", "Yihang Lou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "408c8e4bc46072641819f7ab1d97ba98604c06ba7e91049b8d0e29e1a8ff2db1", "submitted_at": "2026-06-10T08:56:27Z", "updated_at": "2026-08-17T02:54:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.12485v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.12485", "identifiers": {"arxiv": "2606.12485"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.12485"}], "name": "Speculative Rollback Correction for Quality-Diverse Web Agent Imitation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Speculative Rollback Correction lets a student web agent execute short branches before teacher review localizes harmful deviations. Verified continuations and localized corrections feed supervised training while an archive retains diverse successful solutions.", "topics": ["imitation-learning", "data-generation", "robustness"], "updated_at": "2026-08-17", "url": "https://arxiv.org/abs/2606.12485"}
{"aliases": [], "authors": ["Siyu Shen", "Fenghao Xu", "Wenrui Diao", "Kehuan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fed1c6e109cc5d34aebe3bb333955e8bf792264d4c1ed0893d76534a77649afd", "submitted_at": "2026-06-10T20:48:39Z", "updated_at": "2026-06-16T03:15:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.12666v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.12666", "identifiers": {"arxiv": "2606.12666"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.12666"}], "name": "CAPED: Context-Aware Privacy Exposure Defense for Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "CAPED filters mobile screenshots before upload by combining task requirements, screen context and parsed UI elements to mask incidental private content. AndroidWorld and a controlled seeded privacy evaluation examine leakage reduction alongside task utility and prototype costs.", "topics": ["safety-security", "screen-understanding", "mobile-navigation"], "updated_at": "2026-06-16", "url": "https://arxiv.org/abs/2606.12666"}
{"aliases": [], "authors": ["Tianyu Liu", "Allen Xin Wang", "Antonia Panescu", "Lisa Xinyi Chen", "Wenxin Long", "Xinyu Wei", "Yueqian Jing", "Ziyao Zeng", "Jihang Chen", "Sihan Jiang", "Ziqing Wang", "Siyi Gu", "Siyu Chen", "Xinyang Hu", "Haoran Shao", "Leqi Xu", "Wangjie Zheng", "Zhiyuan Cao", "Ada Fang", "Botao Yu", "Kunyang Sun", "Rex Ying", "Arman Cohan", "Qingyu Chen", "Lingzhou Xue", "Kaize Ding", "Yuanqi Du", "Wengong Jin", "Zhuoran Yang", "Marinka Zitnik", "James Zou", "Hua Xu", "Hongyu Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f9506d656ce62661f6d504b7ca1a20a31f59136544dcbbe95d5c7e13a5e8ef31", "submitted_at": "2026-06-10T22:55:30Z", "updated_at": "2026-06-10T22:55:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.12736v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.12736", "identifiers": {"arxiv": "2606.12736"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.12736"}], "name": "Benchmarking AI Agents for Addressing Scientific Challenges Across Scales", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-10", "url": "https://arxiv.org/abs/2606.12736"}
{"aliases": [], "authors": ["Jiaxin Ai", "Tao Hu", "Xuemeng Yang", "Shu Zou", "Hairong Zhang", "Daocheng Fu", "Yu Yang", "Hongbin Zhou", "Nianchen Deng", "Pinlong Cai", "Zhongyuan Wang", "Botian Shi", "Kaipeng Zhang", "Licheng Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "edf7e63803efba89551aa297bff1da9b0babf8ec898a2971fa6c8a20cc58dea8", "submitted_at": "2026-06-11T11:53:32Z", "updated_at": "2026-06-30T03:31:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.13239v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.13239", "identifiers": {"arxiv": "2606.13239"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.13239"}], "name": "ComAct: Reframing Professional Software Manipulation via COM-as-Action Paradigm", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-11", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "ComAct proposes COM program synthesis as an alternative interface for manipulating professional software and introduces ComCADBench with GUI-based comparison agents. ComActor adds a staged training and correction approach, supported by a Windows-container training platform.", "topics": ["automation", "evaluation", "tool-integration", "science"], "updated_at": "2026-06-30", "url": "https://arxiv.org/abs/2606.13239"}
{"aliases": [], "authors": ["Tao Hu", "Jiaxin Ai", "Licheng Wen", "Xueheng Li", "Shu Zou", "Siqi Li", "Nianchen Deng", "Xinyu Cai", "Hongbin Zhou", "Pinlong Cai", "Daocheng Fu", "Yu Yang", "Hairong Zhang", "Botian Shi", "Xuemeng Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4132132662275a598a66c629ca1545f52ab6418767250bc071bc6da48437b223", "submitted_at": "2026-06-11T13:53:21Z", "updated_at": "2026-08-31T10:29:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.13368v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.13368", "identifiers": {"arxiv": "2606.13368"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.13368"}], "name": "IterCAD: An Iterative Multimodal Agent for Visually-Grounded CAD Generation and Editing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2606.13368"}
{"aliases": [], "authors": ["Zihao Wang", "Yiming Li", "Yutong Wu", "Kangjie Chen", "Zheyu Liu", "Fok Kar Wai", "Pin-Yu Chen", "Vrizlynn L. L. Thing", "Bo Li", "Dacheng Tao", "Tianwei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88b7e2ee8717661e4f8b417916e35b844171ecd8a8a2add657cf341c8f759536", "submitted_at": "2026-06-11T14:12:43Z", "updated_at": "2026-07-27T14:19:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.13385v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.13385", "identifiers": {"arxiv": "2606.13385"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.13385"}], "name": "Who Pays the Price? Stakeholder-Centric Prompt Injection Benchmarking for Real-world Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "StakeBench evaluates prompt-injection harm to different stakeholders in shopping web-agent workflows. Its executable cases separate attack objectives and assess both outcomes and execution processes to reveal how the same attack pattern can affect different parties.", "topics": ["evaluation", "safety-security", "web-navigation"], "updated_at": "2026-07-27", "url": "https://arxiv.org/abs/2606.13385"}
{"aliases": [], "authors": ["Baochang Ren", "Xinjie Liu", "Xi Chen", "Yanshuo Liu", "Chenxi Li", "Daqi Gao", "Zeqin Su", "Jintao Xing", "Zirui Xue", "Rui Li", "Xiangyu Zhao", "Shuofei Qiao", "Minting Pan", "Wangmeng Zuo", "Lei Bai", "Dongzhan Zhou", "Ningyu Zhang", "Huajun Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4830ed7fcf26fc506e4ccba4d9dacc3d4b7219cab7a2ece968d06c49c783dc2", "submitted_at": "2026-06-11T17:03:53Z", "updated_at": "2026-06-15T17:50:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.13578v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.13578", "identifiers": {"arxiv": "2606.13578"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.13578"}], "name": "LabVLA: Grounding Vision-Language-Action Models in Scientific Laboratories", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-15", "url": "https://arxiv.org/abs/2606.13578"}
{"aliases": [], "authors": ["Zongsheng Cao", "Bihao Zhan", "Jinxin Shi", "Jiong Wang", "Fangchen Yu", "Zhijie Zhong", "Yingnan Han", "Zijie Guo", "Tianshuo Peng", "Zhuo Liu", "Yi Xie", "Xiang Zhuang", "Shengji Tang", "Yue Fan", "Runmin Ma", "Shiyang Feng", "Xiangchao Yan", "Anran Liu", "Peng Ye", "Wenlong Zhang", "Xiaosong Wang", "Shufei Zhang", "Chunfeng Song", "Fenghua Ling", "Jie Zhou", "Liang He", "Bo Zhang", "Lei Bai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b42fd8e24207e368c2c754e77f6d1cc2c8455203acd995ccf1d872cf2e0c7956", "submitted_at": "2026-06-11T17:58:35Z", "updated_at": "2026-07-17T03:48:22Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.13669v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.13669", "identifiers": {"arxiv": "2606.13669"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.13669"}], "name": "Agents-K1: Towards Agent-native Knowledge Orchestration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-17", "url": "https://arxiv.org/abs/2606.13669"}
{"aliases": [], "authors": ["Zijing Shi", "Meng Fang", "Ling Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73b6da2bbce19e91f4aedbabb7544051ac898c78aad3069cbce71b3c2b44865a", "submitted_at": "2026-04-26T17:46:28Z", "updated_at": "2026-04-26T17:46:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.13686v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.13686", "identifiers": {"arxiv": "2606.13686"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.13686"}], "name": "Benchmarking Web Agent Safety under E-commerce Deceptive Interfaces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-04-26", "url": "https://arxiv.org/abs/2606.13686"}
{"aliases": [], "authors": ["Tejas Agrawal", "Vu Le", "Sumit Gulwani", "Gust Verbruggen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67c586b91361e9ef71fc11531865f4bb97f3207d68eeefee4905a2db14d9e984", "submitted_at": "2026-06-11T18:16:37Z", "updated_at": "2026-06-11T18:16:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.13802v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.13802", "identifiers": {"arxiv": "2606.13802"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.13802"}], "name": "A Benchmark and Framework for Evaluating Next Action Predictions in Spreadsheets", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-11", "url": "https://arxiv.org/abs/2606.13802"}
{"aliases": [], "authors": ["Haoran You", "Yotam Nitzan", "Lingzhi Zhang", "Yifan Gong", "Mang-Tik Chiu", "Connelly Barnes", "Yan Kang", "Yuqian Zhou", "Haitian Zheng", "Eli Shechtman", "Sohrab Amirghodsi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "30c7c0bb8fe80b0c0ba368d2c376b32eef596f8f516fa65e7d244b175b704acb", "submitted_at": "2026-06-11T20:45:26Z", "updated_at": "2026-10-10T17:55:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.13898v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.13898", "identifiers": {"arxiv": "2606.13898"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.13898"}], "name": "HiLo-Token: Input-Adaptive High-Low Frequency Token Compression for Efficient Image Editing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-10", "url": "https://arxiv.org/abs/2606.13898"}
{"aliases": [], "authors": ["Seoyoung Choi", "Minseok Ko", "Hyunseok Lee", "Kunwoong Kim", "Woomin Song", "Chanseok Jeon", "Jinwoo Shin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1fb87da27071ab21519599b7d16359a38118aa358f620de94e66b15cad040828", "submitted_at": "2026-06-12T04:39:01Z", "updated_at": "2026-06-12T04:39:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14106v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.14106", "identifiers": {"arxiv": "2606.14106"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.14106"}], "name": "Naive Visual Memory is Not Enough: A Failure-Mode Study of GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study finds that full-image visual memory can reduce some state errors while worsening action-level failures. Its AGMem alternative stores local crops associated with successful actions or recovery and evaluates their use in desktop GUI tasks.", "topics": ["memory", "interaction-history", "grounding", "robustness", "evaluation"], "updated_at": "2026-06-12", "url": "https://arxiv.org/abs/2606.14106"}
{"aliases": [], "authors": ["Mykola Vysotskyi", "Runqi Lin", "Grzegorz Biziel", "Michal Zakrzewski", "Sebastian Montagna", "Damian Rynczak", "Shreyansh Padarha", "Kumail Alhamoud", "Zihao Fu", "William Lugoloobi", "Kai Rawal", "Hanna Yershova", "Taras Rumezhak", "Guohao Li", "Fazl Barez", "Baoyuan Wu", "Arkadiusz Drohomirecki", "Chris Russell", "Christopher Summerfield", "Adam Mahdi", "Volodymyr Karpiv", "Philip Torr", "Adel Bibi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3eb3a61ae37a45c1415f2da816c8e1458eb02df5bc30c52c1b1bfa46846c5f17", "submitted_at": "2026-06-12T12:32:24Z", "updated_at": "2026-09-28T11:24:26Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2606.14397v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.14397", "identifiers": {"arxiv": "2606.14397"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.14397"}], "name": "Running the Gauntlet: Hard Agentic Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2606.14397"}
{"aliases": [], "authors": ["Xinyu Qiu", "Yunzhu Zhang", "Heng Jia", "Shuheng Shen", "Changhua Meng", "Linchao Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a5e949a760f236c08ae3cebc32d8a52220b1101f5324770d40bd7d054351f6a", "submitted_at": "2026-06-12T15:58:18Z", "updated_at": "2026-06-12T15:58:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14579v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.14579", "identifiers": {"arxiv": "2606.14579"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.14579"}], "name": "VISTA: View-Consistent Self-Verified Training for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "VISTA forms GRPO comparison groups from target-preserving crops of the same GUI instance with exactly remapped target boxes. A conditional, advantage-weighted cross-view anchor supplies additional supervision without entering the group baseline, and evaluation includes view-consistency robustness.", "topics": ["grounding", "reinforcement-learning", "robustness", "data-generation"], "updated_at": "2026-06-12", "url": "https://arxiv.org/abs/2606.14579"}
{"aliases": [], "authors": ["Chenxin Li", "Zhengyao Fang", "Zhengyang Tang", "Pengyuan Lyu", "Xingran Zhou", "Xin Lai", "Fei Tang", "Liang Wu", "Yiduo Guo", "Weinong Wang", "Junyi Li", "Yi Zhang", "Yang Ding", "Huawen Shen", "Sunqi Fan", "Shangpin Peng", "Zheng Ruan", "Anran Zhang", "Benyou Wang", "Chengquan Zhang", "Han Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "280ec12bcb246f1f07cf3627011945e8a9ff4afed93b1fb2937f2d6b1bc1323b", "submitted_at": "2026-06-12T15:01:32Z", "updated_at": "2026-06-12T15:01:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14832v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.14832", "identifiers": {"arxiv": "2606.14832"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.14832"}], "name": "PhoneHarness: Harnessing Phone-Use Agents through Mixed GUI, CLI, and Tool Actions", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "PhoneHarness routes phone workflows across GUI actions, device-side commands and host tools while retaining auditable execution traces. Its accompanying benchmark checks observable task side effects, separating mixed-action execution from plausible final-answer generation.", "topics": ["mobile-navigation", "tool-integration", "evaluation", "workflow"], "updated_at": "2026-06-12", "url": "https://arxiv.org/abs/2606.14832"}
{"aliases": [], "authors": ["Sina Hajimiri", "Masih Aminbeidokhti", "Jose Dolz", "Ismail Ben Ayed", "Issam H. Laradji", "Spandana Gella", "Nicolas Gontier"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50f1acd5727f84946cf95d181f270dbc4f46f15d24339e6cb257fea1b23b21a0", "submitted_at": "2026-06-12T23:30:14Z", "updated_at": "2026-08-30T17:49:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.15017v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.15017", "identifiers": {"arxiv": "2606.15017"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.15017"}], "name": "Are Online Skill and Memory Modules Always Worth Their Tokens? A Budget-Constrained Study of Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares online skill and memory augmentations with a vanilla web agent given the same total inference-token budget. Across WebArena and WorkArena settings, it studies when module overhead removes aggregate gains and how rerun variability changes conclusions.", "topics": ["memory", "efficiency", "evaluation", "robustness"], "updated_at": "2026-08-30", "url": "https://arxiv.org/abs/2606.15017"}
{"aliases": [], "authors": ["Mina Mohammadmirzaei", "Jeffrey Flanigan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7bf0297fba942441b19bcfeba147838bdef888b428f1150c2462d8bf3d986882", "submitted_at": "2026-06-13T00:32:24Z", "updated_at": "2026-10-09T19:45:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.15034v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.15034", "identifiers": {"arxiv": "2606.15034"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.15034"}], "name": "OSGuard: A Benchmark for Safety in Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates action-level guardrails and full computer tasks with environment-dependent safety constraints, separating safe completion from nominal success.", "topics": ["safety-security", "evaluation"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2606.15034"}
{"aliases": [], "authors": ["Xinhang Ma", "Taoran Li", "Chaowei Xiao", "Zhiyuan Yu", "Ning Zhang", "Yevgeniy Vorobeychik"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "baa1639a71629640fef60ca1d654a43c23e500f57047f8061219d6355ad8aca8", "submitted_at": "2026-06-13T02:09:08Z", "updated_at": "2026-09-28T04:15:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.15057v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.15057", "identifiers": {"arxiv": "2606.15057"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.15057"}], "name": "AutoDojo: A Generative Benchmark for Evaluating Prompt Injection Defenses in LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2606.15057"}
{"aliases": [], "authors": ["Yuxin Zhang", "Ju Fan", "Meihao Fan", "Shaolei Zhang", "Xiaoyong Du"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "84fb4056dbd651bb6fcfb2661b25634d6526913471da92750f193fcd7e513e56", "submitted_at": "2026-06-13T13:46:47Z", "updated_at": "2026-06-13T13:46:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.15300v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.15300", "identifiers": {"arxiv": "2606.15300"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.15300"}], "name": "CODA-BENCH: Can Code Agents Handle Data-Intensive Tasks?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-13", "url": "https://arxiv.org/abs/2606.15300"}
{"aliases": [], "authors": ["Lipeng He", "Yihan Wang", "Jiawen Zhang", "N. Asokan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "753c6b7737dff3814d7ae26a3c14f30cc7af02aa957ca5ec80a1663400629c8f", "submitted_at": "2026-06-13T19:15:44Z", "updated_at": "2026-06-13T19:15:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.15441v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.15441", "identifiers": {"arxiv": "2606.15441"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.15441"}], "name": "Defending against Adaptive Prompt Injection Attacks via Reasoning-enabled Task Alignment", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-13", "url": "https://arxiv.org/abs/2606.15441"}
{"aliases": [], "authors": ["Jiwan Chung", "JiHyuk Byun", "Vibhav Vineet", "Seon Joo Kim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67454ecc0b32b52c65be8b2c37c32f73d793750f6c14d9cb5a7aecf6f094d85b", "submitted_at": "2026-04-08T07:25:29Z", "updated_at": "2026-08-04T05:22:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.15673v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.15673", "identifiers": {"arxiv": "2606.15673"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.15673"}], "name": "Where Did It Go Wrong? Process-Level Evaluation of Web Agents with Semantic State Tracking", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-04-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2606.15673"}
{"aliases": [], "authors": ["Y. H. Zhou", "Z. M. Ma", "Y. J. Zhou", "Y. T. Li", "H. X. Xiang", "Y. M. Cheng", "T. L. Chen", "K. J. Zhang", "Z. H. Nan", "J. H. Ni", "Z. Wu", "Q. Y. Pan", "S. Zhang", "S. Cheng", "M. Y. Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6311b22ddb2262c74efbe79acc8856335f6870aa24197c507f3f7b048d6f7901", "submitted_at": "2026-06-15T12:53:55Z", "updated_at": "2026-06-15T12:53:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.16659v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.16659", "identifiers": {"arxiv": "2606.16659"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.16659"}], "name": "FraudSMSWalker: Benchmarking Agentic Large Language Models for SMS-to-Webpage Fraud Detection", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "FraudSMSWalker tests browser agents on bilingual SMS-to-webpage fraud judgments while hiding URL and reputation shortcuts. Its controlled chains include difficult benign login/payment cases and evaluate whether classifications remain supported by the visible message and page evidence.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-06-15", "url": "https://arxiv.org/abs/2606.16659"}
{"aliases": [], "authors": ["Lawrence Keunho Jang", "Andrew Keunwoo Jang", "Jing Yu Koh", "Ruslan Salakhutdinov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "acc529f96d6df52e3560f66df8e0630dabdae3ab30cadc8b59123f347eba187a", "submitted_at": "2026-06-15T14:08:09Z", "updated_at": "2026-06-15T14:08:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.16748v1"}}, "environments": ["desktop", "linux", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.16748", "identifiers": {"arxiv": "2606.16748"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.16748"}], "name": "MyPCBench: A Benchmark for Personally Intelligent Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "MyPCBench tests personal-assistant tasks on a Linux desktop with simulated web applications and data for a single fictional persona. A shared computer-and-shell interface enables evaluation of cross-application workflows that depend on personal context and account-like state.", "topics": ["evaluation", "cross-app", "workflow"], "updated_at": "2026-06-15", "url": "https://arxiv.org/abs/2606.16748"}
{"aliases": [], "authors": ["Anqi Zou", "Han Deng", "Chengyu Zhang", "Junquan Hu", "Yu Wang", "Yuxiang Xing", "Aokai Zhang", "Hanling Zhang", "Zhaoyang Liu", "Ben Fei", "Zhihui Wang", "Wanli Ouyang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7487f31097bc8de90dfbe836051d47f407c26e31a42bd751bedfd38cf4f1742c", "submitted_at": "2026-06-15T14:42:33Z", "updated_at": "2026-06-15T14:42:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.16802v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.16802", "identifiers": {"arxiv": "2606.16802"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.16802"}], "name": "LabOSBench: Benchmarking Computer Use Agents for Scientific Instrument Control", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "LabOSBench evaluates GUI agents on browser-based scientific instrument simulators with configurable tasks and execution-based checks. Its subtask and end-to-end tests examine feedback-driven adjustment and long workflows without controlling physical instruments.", "topics": ["evaluation", "science", "long-horizon"], "updated_at": "2026-06-15", "url": "https://arxiv.org/abs/2606.16802"}
{"aliases": [], "authors": ["Ye Jin", "Yangyang Xu", "Jun Zhu", "Yibo Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "71a018082287aec9d479e6f6c093dac168465186278bcd1248fb05f33867a199", "submitted_at": "2026-06-15T18:02:55Z", "updated_at": "2026-06-15T18:02:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.17162v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.17162", "identifiers": {"arxiv": "2606.17162"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.17162"}], "name": "MemSlides: A Hierarchical Memory Driven Agent Framework for Personalized Slide Generation with Multi-turn Local Revision", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-15", "url": "https://arxiv.org/abs/2606.17162"}
{"aliases": [], "authors": ["Jaehun Jung", "Ximing Lu", "Brandon Cui", "Muhammad Khalifa", "Shaokun Zhang", "Hao Zhang", "Jin Xu", "Amala Sanjay Deshmukh", "Karan Sapra", "Andrew Tao", "Yejin Choi", "Jan Kautz", "Mingjie Liu", "Yi Dong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d7e6e0b66f2923cb66ab7aa404bedf9f4e1c73ae00df2e9fbfa7a5c3795cac4c", "submitted_at": "2026-06-15T22:04:11Z", "updated_at": "2026-06-15T22:04:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.17321v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.17321", "identifiers": {"arxiv": "2606.17321"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.17321"}], "name": "ProCUA-SFT Technical Report", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "ProCUA-SFT generates tasks and synthetic trajectories on live desktops seeded with document content, checking task preconditions before rollout. The resulting step-prefix examples are formatted to match inference context and used for supervised GUI-agent fine-tuning.", "topics": ["data-generation", "imitation-learning", "evaluation"], "updated_at": "2026-06-15", "url": "https://arxiv.org/abs/2606.17321"}
{"aliases": [], "authors": ["Shiqi He", "Yue Cui", "Feijie Wu", "Xinyu Ma", "Jiaheng Lu", "Yaliang Li", "Bolin Ding", "Mosharaf Chowdhury"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81e6d5fdedfa7a16876f2a56f4f24a557562cedaffd2978c6fca994d9989685e", "submitted_at": "2026-06-16T08:04:35Z", "updated_at": "2026-06-16T08:04:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.17645v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.17645", "identifiers": {"arxiv": "2606.17645"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.17645"}], "name": "Beyond Domains: Reusing Web Skills via Transferable Interaction Patterns", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-16", "url": "https://arxiv.org/abs/2606.17645"}
{"aliases": [], "authors": ["Bojie Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "19fd8d8363ef6a7c58e42309f25dea1b9c5b6bca09b5d823f8a7cc7a80d61f0f", "submitted_at": "2026-06-16T13:40:21Z", "updated_at": "2026-06-16T13:40:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.17929v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.17929", "identifiers": {"arxiv": "2606.17929"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.17929"}], "name": "PreAct: Computer-Using Agents that Get Faster on Repeated Tasks", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "PreAct compiles successful screen-interaction runs into guarded state-machine programs for repeated tasks. It validates programs from clean states before storage and returns control to the agent when replayed screen expectations fail, with tests across mobile, desktop and web.", "topics": ["recording", "memory", "efficiency", "automation"], "updated_at": "2026-06-16", "url": "https://arxiv.org/abs/2606.17929"}
{"aliases": [], "authors": ["Jingyuan Huang", "Zuming Huang", "Yucheng Shi", "Tianze Yang", "Xiaoming Zhai", "Wei Chu", "Ninghao Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49b935d8993be0d03ae778375f908b68b3ec3c614a7406b6b3f6a00f2a89ef1f", "submitted_at": "2026-06-16T16:02:41Z", "updated_at": "2026-06-17T08:58:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.18101v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.18101", "identifiers": {"arxiv": "2606.18101"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.18101"}], "name": "Trust the Right Teacher: Quality-Aware Self-Distillation for GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes quality-aware self-distillation for GUI coordinate generation under student-produced prefixes. A correctness-aware gate suppresses coordinate-token supervision that cannot reach the target box, while teacher-probability scaling adjusts the remaining signal strength.", "topics": ["grounding", "imitation-learning", "evaluation"], "updated_at": "2026-06-17", "url": "https://arxiv.org/abs/2606.18101"}
{"aliases": [], "authors": ["Komal Thareja", "Hamza Safri", "Rajiv Mayani", "Anirban Mandal", "Ewa Deelman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a403da12d9dc7731a20830387c857dc7b6bbb4ec0685fee85094ab2b950baba", "submitted_at": "2026-06-16T19:21:09Z", "updated_at": "2026-09-08T15:30:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.18425v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.18425", "identifiers": {"arxiv": "2606.18425"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.18425"}], "name": "Complexity and Scale in AI-Assisted Workflow Management: A Federated Learning Case Study", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-08", "url": "https://arxiv.org/abs/2606.18425"}
{"aliases": [], "authors": ["Ziyan Jiang", "Li An", "Yujian Liu", "Jiabao Ji", "Qiucheng Wu", "Jacob Andreas", "Yang Zhang", "Shiyu Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2605cedb4359aa0cf1e694455903871ed334dd37b7b8e15edf31bac9d5e51b13", "submitted_at": "2026-06-16T19:57:07Z", "updated_at": "2026-06-16T19:57:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.18448v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.18448", "identifiers": {"arxiv": "2606.18448"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.18448"}], "name": "VISUALSKILL: Multimodal Skills for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "VISUALSKILL organizes application-specific text and figures into hierarchical skills retrieved on demand through an MCP tool. Skills combine authored documentation with live UI exploration, and matched multimodal/text-only comparisons test their utility for interface identification and workflow-state verification.", "topics": ["memory", "tool-integration", "screen-understanding", "long-horizon"], "updated_at": "2026-06-16", "url": "https://arxiv.org/abs/2606.18448"}
{"aliases": [], "authors": ["Yujin Zhang", "Daye Nam"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c52695e860206d02735839853416965ec3d42fac134ebfa0e8f2f5f5ea043914", "submitted_at": "2026-06-17T04:13:20Z", "updated_at": "2026-06-17T04:13:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.18671v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.18671", "identifiers": {"arxiv": "2606.18671"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.18671"}], "name": "HANSEL: Extracting Breadcrumbs from Web Agent Trajectories for Interactive Verification", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "HANSEL extracts evidence pages and snippets from browser-agent traces and preserves relevant page state in interactive verification views. It flags untraceable answers and evaluates evidence selection and human verification effort separately.", "topics": ["human-in-the-loop", "observability", "recording", "evaluation"], "updated_at": "2026-06-17", "url": "https://arxiv.org/abs/2606.18671"}
{"aliases": [], "authors": ["Zhimin Fan", "Hongwei Yu", "Yeqing Shen", "Haolong Yan", "Guozhen Peng", "Tianhao Peng", "Yudong Zhang", "Xiaowen Zhang", "Kaijun Tan", "Zheng Ge", "Xiangyu Zhang", "Daxin Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1929dfb5074e159d83006c7bbb470237a0c557e0723ac631116aa5e0369f9247", "submitted_at": "2026-06-17T10:07:51Z", "updated_at": "2026-06-17T10:07:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.18890v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.18890", "identifiers": {"arxiv": "2606.18890"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.18890"}], "name": "Skill-Guided Continuation Distillation for GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Skill-Guided Continuation Distillation first lets a policy reach its own off-trajectory states, then uses skill guidance to collect successful continuations. Mixing those continuations with expert traces supplies supervision for recovery states absent from demonstrations.", "topics": ["imitation-learning", "data-generation", "robustness", "planning"], "updated_at": "2026-06-17", "url": "https://arxiv.org/abs/2606.18890"}
{"aliases": [], "authors": ["Yingshan Susan Wang", "Cedegao E. Zhang", "Linlu Qiu", "Zexue He", "Pengyuan Li", "Alex Pentland", "Roger P. Levy", "Yoon Kim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae39265a336636c79c67b28d0d79861e81e8eb9af18efdc9a0bf6df7553d3ca4", "submitted_at": "2026-06-17T17:58:48Z", "updated_at": "2026-06-25T03:03:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.19336v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.19336", "identifiers": {"arxiv": "2606.19336"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.19336"}], "name": "Learning User Simulators with Turing Rewards", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-25", "url": "https://arxiv.org/abs/2606.19336"}
{"aliases": [], "authors": ["Li Gu", "Zihuan Jiang", "Linqiang Guo", "Zhixiang Chi", "Ziqiang Wang", "Huan Liu", "Yuanhao Yu", "Tse-Hsun Chen", "Yang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9bf8b07e08b275d90fcbfc63d867f54f5579509132cc9cf694d01447b7f91b78", "submitted_at": "2026-06-16T02:36:22Z", "updated_at": "2026-06-16T02:36:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.19388v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.19388", "identifiers": {"arxiv": "2606.19388"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.19388"}], "name": "Beyond the GUI Paradigm: Do Mobile Agents Need the Phone Screen?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-16", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "This study compares general coding agents using mobile command-line access with reproducible GUI agents on AndroidWorld and MobileWorld. It also supplies oracle CLI analyses and a separate task suite for operations such as bulk changes, filtering and hidden device state.", "topics": ["evaluation", "automation", "cross-app", "tool-integration"], "updated_at": "2026-06-16", "url": "https://arxiv.org/abs/2606.19388"}
{"aliases": [], "authors": ["Fukang Wen", "Zhijie Wang", "Ruilin Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e1ecab5a6f5e4aa5c0d0f8ee9d9d426c5fa480c9c768b4c97fbdc50ac1bf669", "submitted_at": "2026-06-17T15:19:19Z", "updated_at": "2026-06-17T15:19:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.19409v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.19409", "identifiers": {"arxiv": "2606.19409"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.19409"}], "name": "OpenRath: Session-Centered Runtime State for Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-17", "url": "https://arxiv.org/abs/2606.19409"}
{"aliases": [], "authors": ["Dhaval C. Patel", "Kaoutar El Maghraoui", "Shuxin Lin", "Yusheng Li", "Tianjun Feng", "Chun-Yi Tsai", "Yihan Sun", "Wei Alexander Xin", "Akshat Bhandari", "Tanisha Rathod", "Aaron Fan", "Sanskruti Vijay Shejwal", "Tomas Pasiecznik", "Sagar Chethan Kumar", "Tanmay Agarwal", "Rohith Kanathur", "Sam Colman", "Amaan Sheikh", "Dev Bahl", "Ann Li", "Krish Veera", "Alimurtaza Mustafa Merchant", "Shambhawi Baswaraj Bhure", "Sajal Kumar Goyla", "Chengrui Li", "Kirthana Natarajan", "Rui Li", "Thomas Ajai", "Rujing Li", "Vivek G. Iyer", "Sanjaii Vijayakumar", "Yitong Bai", "Ayal Yakobe", "Darief Maes", "Yassine Jebbouri", "Tianyang Xu", "Thai Quoc On", "Vera Mazeeva", "Winston Li", "Yuval Shemla", "Yeshitha Bhuvanesh", "Rushin Bhatt", "Siddharth Chethan Gowda", "Alisha Vinod", "Caroline Cahill", "Shriya Aishani Rachakonda", "Yunfeng Chen", "Aryaman Agrawal", "Aman Upganlawar", "Mao Le Jonathan Ang", "Yubin Sally Go", "Madhav Rajkondawar", "Yang-Jung Chen", "Trisha Maturi", "Ananya Kapoor", "Andrew Li", "Shrey Arora", "Mana Abbaszadeh", "Shen Li", "Charles Xu", "Byeolah Kwon"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61d771e3b0df453da3ba6d1c3664ae8759beda75c7eb40bca4b899d40421d48f", "submitted_at": "2026-06-18T02:02:46Z", "updated_at": "2026-06-18T02:02:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.19704v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.19704", "identifiers": {"arxiv": "2606.19704"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.19704"}], "name": "Beyond Static Leaderboards: Predictive Validity for the Evaluation of LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-18", "url": "https://arxiv.org/abs/2606.19704"}
{"aliases": [], "authors": ["Guangyi Liu", "Gao Wu", "Congxiao Liu", "Pengxiang Zhao", "Liang Liu", "Mading Li", "Qi Zhang", "Mengyan Wang", "Liang Guo", "Yong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cbbb1bd57057e465ce9d991d3f6e2bb52112fc133a5846fc8cb6ce51eddc2121", "submitted_at": "2026-06-18T08:26:09Z", "updated_at": "2026-06-18T08:26:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.19926v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.19926", "identifiers": {"arxiv": "2606.19926"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.19926"}], "name": "MemGUI-Agent: An End-to-End Long-Horizon Mobile GUI Agent with Proactive Context Management", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "MemGUI-Agent makes context management an action chosen by the same policy that controls the mobile UI. Structured history, UI-state and recent-step fields are trained from annotated trajectories to retain cross-app facts while limiting context growth.", "topics": ["memory", "interaction-history", "long-horizon", "imitation-learning", "data-generation"], "updated_at": "2026-06-18", "url": "https://arxiv.org/abs/2606.19926"}
{"aliases": [], "authors": ["Guangyi Liu", "Pengxiang Zhao", "Gao Wu", "Yiwen Yin", "Mading Li", "Liang Liu", "Congxiao Liu", "Zhang Qi", "Mengyan Wang", "Liang Guo", "Jiangning Zhang", "Yong Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cab809618ea4e218da26b4ae11bf937859c3a4bc240c91dd27b3a0d333737069", "submitted_at": "2026-06-18T08:29:33Z", "updated_at": "2026-08-07T03:54:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.19930v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.19930", "identifiers": {"arxiv": "2606.19930"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.19930"}], "name": "MobileForge: Annotation-Free Adaptation for Mobile GUI Agents with Hierarchical Feedback-Guided Policy Optimization", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileForge couples real-app exploration and generated tasks with hierarchical feedback for mobile-agent adaptation. Its optimizer turns trajectory outcomes, step feedback and corrective hints into contextualized GRPO updates, with evaluation on AndroidWorld and a GUI-only MobileWorld split.", "topics": ["mobile-navigation", "reinforcement-learning", "data-generation", "evaluation"], "updated_at": "2026-08-07", "url": "https://arxiv.org/abs/2606.19930"}
{"aliases": [], "authors": ["Yuexing Hao", "Xiaomin Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b370491b37d5bf25c624f2bac654e438459c8084b5da292fb18a0bf9142adaa2", "submitted_at": "2026-06-18T15:25:42Z", "updated_at": "2026-06-18T15:25:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.20363v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.20363", "identifiers": {"arxiv": "2606.20363"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.20363"}], "name": "Automating SKILL.md Generation for Computer-Using Agents via Interaction Trajectory Mining", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Segments and clusters GUI trajectories into inspectable skills, then trains a skill-aware policy from those annotations. Its diagnostic results distinguish readable source-domain clusters from reliable transfer and report limited or absent downstream gains.", "topics": ["memory", "data-generation", "reinforcement-learning", "evaluation"], "updated_at": "2026-06-18", "url": "https://arxiv.org/abs/2606.20363"}
{"aliases": [], "authors": ["Daniel Beechey", "Derek Yuen", "Jianheng Liu", "Dezhao Luo", "Tiantian He", "Weilin Luo", "Jun Wang", "Kun Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e761b20caa6882605ecfd6cd001aeda4a2d36b2e702eea9a6d70451fc0ea9eb", "submitted_at": "2026-05-26T13:32:14Z", "updated_at": "2026-05-26T13:32:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.20622v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:51.952206+00:00", "role": "primary", "sha256": "5cd7fb95d5ee8e8c473e0c9d5b892bf50ee425604c6a320e8e9a12eff17120a1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:51.952206+00:00", "role": "primary", "sha256": "5cd7fb95d5ee8e8c473e0c9d5b892bf50ee425604c6a320e8e9a12eff17120a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:00.181958+00:00", "role": "primary", "sha256": "e4711466886372760daa83384cb4d129d54fdcfe320be6e91fb7b4971d1a61de", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2606.20622v1"}], "id": "arxiv:2606.20622", "identifiers": {"arxiv": "2606.20622"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:51.952206+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.20622"}], "name": "Darwin Mobile Agent: A Roadmap for Self-Evolution", "notes": "The implemented stage uses selected tasks, an external judge and heuristic history management. Autonomous curricula, verification and memory are roadmap directions, not completed capabilities established by this record. No runtime or results were reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (2 Agent-Environment Interface and Policy Optimisation; 3 Scalable Asynchronous Mobile Environment; 3.2 Persistent State Management; 3.3 Task Lifecycle; 4.2 Rollout and Optimisation; 4.3 Anatomy of the Self-Evolving Agent); not an entire-paper reading or reproduction claim.", "published_at": "2026-05-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes asynchronous reinforcement-learning infrastructure using parallel cloud-hosted Android instances, explicit task lifecycles, and persistent-state management for mobile GUI interaction.", "topics": ["reinforcement-learning", "mobile-navigation", "evaluation"], "updated_at": "2026-05-26", "url": "https://arxiv.org/abs/2606.20622"}
{"aliases": [], "authors": ["Yurun Chen", "Biao Yi", "Keting Yin", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "735799024784157d28a03727b9d39b5c21ef747f6978b657e2fa793fe9f39e1e", "submitted_at": "2026-06-02T06:33:51Z", "updated_at": "2026-06-02T06:33:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.20636v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.20636", "identifiers": {"arxiv": "2606.20636"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.20636"}], "name": "SkillHarness: Harnessing Safe Skills for Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-02", "url": "https://arxiv.org/abs/2606.20636"}
{"aliases": [], "authors": ["Xuelong Dai", "Jianyu Ma", "Boyang Ma", "Biwei Yan", "Yijun Yang", "Yue Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fea441692131ec677e4a9ae54f9ecef1f31d422e20884b496b124e99392e59c1", "submitted_at": "2026-06-16T15:31:33Z", "updated_at": "2026-10-15T04:29:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.20717v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.20717", "identifiers": {"arxiv": "2606.20717"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.20717"}], "name": "MIRAGE: Stealthy Visual Prompt Injection for Vulnerability Detection in Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "MIRAGE evaluates targeted next-action hijacking from adversarial images restricted to legitimate third-party webpage regions. It combines diffusion-based image generation with execution-aware optimization and tests the constrained visual-injection setting on web-agent frameworks.", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2026-10-15", "url": "https://arxiv.org/abs/2606.20717"}
{"aliases": [], "authors": ["Aagam Sogani", "Botao Rui", "Swetha Vaidyanathan", "Rishi Agarwal", "Minghao Yan", "Shivaram Venkataraman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d9793279a8295c35cc745e2aeb1da499654130f53e145e6835ab03d309440f0d", "submitted_at": "2026-06-16T23:00:25Z", "updated_at": "2026-06-29T04:16:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.20724v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.20724", "identifiers": {"arxiv": "2606.20724"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.20724"}], "name": "When Web Agents Finish but Still Fail: Reproducible Triggers and Trace Diagnostics for Parallel Web Exploration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-29", "url": "https://arxiv.org/abs/2606.20724"}
{"aliases": [], "authors": ["Ahmed Awadallah", "Sahil Gupta", "Yash Lara", "Yadong Lu", "Hussein Mozannar", "Akshay Nambi", "Zach Nussbaum", "Yash Pandya", "Aravind Rajeswaran", "Corby Rosset", "Alexey Taymanov", "Luiz do Valle", "Vibhav Vineet", "Spencer Whitehead", "Andrew Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "414eced85da9f76828e4398b977f9b8dac77236c97d1594108c8f5dfc3be7cbb", "submitted_at": "2026-06-18T17:53:03Z", "updated_at": "2026-07-22T17:47:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.20785v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.20785", "identifiers": {"arxiv": "2606.20785"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.20785"}], "name": "Fara-1.5: Scalable Learning Environments for Computer Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds browser-agent training trajectories with live and synthetic environments, solvers, user simulation, and complementary correctness and efficiency verifiers.", "topics": ["data-generation", "web-navigation", "evaluation"], "updated_at": "2026-07-22", "url": "https://arxiv.org/abs/2606.20785"}
{"aliases": [], "authors": ["Dayeon Kang", "Hyejun Jeong", "Jade Sheffey", "Pubali Datta", "Amir Houmansadr"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b2c6080f0de11c06fe73c30195c7c8529e07de283dc45fa47709072846d385a", "submitted_at": "2026-06-18T20:07:39Z", "updated_at": "2026-06-18T20:07:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.20910v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.20910", "identifiers": {"arxiv": "2606.20910"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.20910"}], "name": "Whose Agent Are You? Multi-Layer Fingerprinting and Attribution of Autonomous Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds a logging framework that combines network/protocol fingerprints with browser interaction behavior to attribute web-agent traffic. A decision-tree classifier is evaluated across six agent implementations, human browsing and conventional crawlers.", "topics": ["safety-security", "observability", "evaluation"], "updated_at": "2026-06-18", "url": "https://arxiv.org/abs/2606.20910"}
{"aliases": [], "authors": ["Sergei Trashchenkov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "85212bbfcc1bf8951e85ed131db04fc4d87728fca3e46a3a3f9fe0054316115e", "submitted_at": "2026-06-18T21:30:39Z", "updated_at": "2026-07-02T14:10:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.20950v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.20950", "identifiers": {"arxiv": "2606.20950"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.20950"}], "name": "Power Systems Agent Benchmark: Executable Evaluation of AI Agents in Electric Power Engineering", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-02", "url": "https://arxiv.org/abs/2606.20950"}
{"aliases": [], "authors": ["Jiancheng Zhao", "Xiang Ji", "Yifan Zhan", "Zunian Wan", "Yinqiang Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "92b68abcb92c6341d663f6e0b9b6f9bafc2d9d16e8ea5a9214fcea161899a1f2", "submitted_at": "2026-06-19T01:56:25Z", "updated_at": "2026-06-19T01:56:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.21033v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.21033", "identifiers": {"arxiv": "2606.21033"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.21033"}], "name": "MoECodec: Image Compression for joint human and machine perception via Mixture-of-Experts", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-19", "url": "https://arxiv.org/abs/2606.21033"}
{"aliases": [], "authors": ["Noe Y. Flandre", "Alexander C. Nwala", "Philippe J. Giabbanelli"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b1786a16e5f159c3bd3ede1cebd6a4dbd4f3aad6e093dfadf4e8e6f8d21a6f2", "submitted_at": "2026-06-19T16:03:53Z", "updated_at": "2026-06-19T16:03:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.21565v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.21565", "identifiers": {"arxiv": "2606.21565"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.21565"}], "name": "Composing Verifiable Conceptual Models via Building Blocks: Towards Design-Time Verification of Agentic AI Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-19", "url": "https://arxiv.org/abs/2606.21565"}
{"aliases": [], "authors": ["Vincent Siu", "Manasi Sharma", "Dawn Song", "Daniel Yue Zhang", "Chenguang Wang", "Ying Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "69e668a9491a31e8d8849bab8cc63eede9878fff10ccb0ae0a211610e90819b3", "submitted_at": "2026-06-19T18:00:14Z", "updated_at": "2026-08-19T17:29:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.21654v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.21654", "identifiers": {"arxiv": "2606.21654"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.21654"}], "name": "ChainWorld: Composing Long-Horizon Desktop Workloads from Atomic OSWorld Tasks", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "ChainWorld composes compatible atomic OSWorld tasks into multi-objective desktop chains while preserving their evaluators. It compares giving the full sequence at once with revealing goals turn by turn, exposing different state and session-management failures.", "topics": ["evaluation", "long-horizon", "state-tracking", "workflow"], "updated_at": "2026-08-19", "url": "https://arxiv.org/abs/2606.21654"}
{"aliases": [], "authors": ["Yikun Fu", "Bowen Fu", "Zhenyu Wu", "Shuang Cheng", "Xiaowei Sun", "Bowen Yang", "Zehao Li", "Yibo Zhao", "Zichen Ding", "Zhoumianze Liu", "Shijie Wang", "Biqing Qi", "Bowen Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6dc52f58bb2237d12041fa8b994b20bd1e41290aaab0c70d2c5dd79fcaca37b6", "submitted_at": "2026-06-21T15:32:57Z", "updated_at": "2026-06-21T15:32:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.22557v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}], "id": "arxiv:2606.22557", "identifiers": {"arxiv": "2606.22557"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.22557"}], "name": "MacAgentBench: Benchmarking AI Agents on Real-World macOS Desktop", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces macOS tasks spanning GUI and command-line interaction, with deterministic checks and intermediate-progress scoring for application workflows.", "topics": ["evaluation", "cross-app", "tool-integration"], "updated_at": "2026-06-21", "url": "https://arxiv.org/abs/2606.22557"}
{"aliases": [], "authors": ["Weidi Luo", "Qiming Zhang", "Yihao Quan", "Mingyu Jin", "Jie Cai", "Chaowei Xiao", "Jingcheng Niu", "Zhen Xiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd97e92e0731e8b00dba47217fa60a3b53a84f19b716c3d4347c761294dd9d32", "submitted_at": "2026-06-21T21:13:55Z", "updated_at": "2026-10-05T17:21:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.22673v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T04:48:34Z", "role": "primary", "sha256": "4b1219a8d3b34b9ecfabab989eed959ab7e2263500ab2034fc56105db1c559eb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.22673%2C2610.17816%2C2610.20597%2C2610.23857%2C2610.25601&max_results=5"}], "id": "arxiv:2606.22673", "identifiers": {"arxiv": "2606.22673"}, "kind": "paper", "last_verified_at": "2026-10-21T04:48:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.22673"}], "name": "Agent MechSuits: Mechanistic Subspace Safety Steering for Multi-Turn CLI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2606.22673"}
{"aliases": [], "authors": ["Yanhang Li", "Zhichao Fan", "Zexin Zhuang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64d3fde34f6900366244901080c8c987da5bec6af51df37a5b03f6de179745ab", "submitted_at": "2026-06-22T05:13:25Z", "updated_at": "2026-06-22T05:13:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.22864v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.22864", "identifiers": {"arxiv": "2606.22864"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.22864"}], "name": "When AUC 0.998 Is Not Enough: A Candidate Evaluation Protocol for Hidden-State Probes of Indirect Prompt Injection in Multimodal Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-22", "url": "https://arxiv.org/abs/2606.22864"}
{"aliases": [], "authors": ["Yincheng Zhou", "Athena Zhuoming Zhong", "Shijie Zhang", "Kevin Zhang", "Teresa Xiaotao Shang", "Shanghang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbd8892e06a5e847c16aa3101b98b3d669a783b84ee8e3d4d4d810754914028b", "submitted_at": "2026-06-22T07:28:55Z", "updated_at": "2026-06-22T07:28:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.22948v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.22948", "identifiers": {"arxiv": "2606.22948"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.22948"}], "name": "ENVS: Environment-Native Verified Search for Long-Horizon GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "ENVS branches over GUI actions in live desktop VMs, verifies successful leaves, and converts them into balanced step-level training supervision. It also introduces OSWorld-Noisy to test recovery from dynamic interruptions while retaining the underlying tasks.", "topics": ["data-generation", "reinforcement-learning", "long-horizon", "robustness", "evaluation"], "updated_at": "2026-06-22", "url": "https://arxiv.org/abs/2606.22948"}
{"aliases": [], "authors": ["Anmol Goel", "Iryna Gurevych"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da8190bc5afc0bae400fc8ab5e7d80a7f34488c3400dce4477ca61635b19d5b7", "submitted_at": "2026-06-22T11:36:58Z", "updated_at": "2026-09-10T18:49:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.23189v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.23189", "identifiers": {"arxiv": "2606.23189"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.23189"}], "name": "Capable but Careless: Do Computer-Use Agents Follow Contextual Integrity?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "AgentCIBench evaluates whether computer-use agents disclose information in ways that violate its original context. Deterministically scored scenarios test nearby sensitive UI content, ambiguous requests and recipient mismatches, including end-to-end agent execution.", "topics": ["evaluation", "safety-security", "cross-app"], "updated_at": "2026-09-10", "url": "https://arxiv.org/abs/2606.23189"}
{"aliases": [], "authors": ["Shanhui Zhao", "Jiacheng Liu", "Guohong Liu", "Jichao Yan", "Jialei Ye", "Yuhao Yang", "Hao Wen", "Shizuo Tian", "Yizhen Yuan", "Yuxuan Chen", "Yunxin Liu", "Ju Ren", "Ya-Qin Zhang", "Chao Huang", "Yao Guo", "Yuanchun Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5615423ca86eb1087e52310434bcf6e1a6f79444da1c83689f68e72ee39f3e5d", "submitted_at": "2026-06-22T15:02:42Z", "updated_at": "2026-06-22T15:02:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.23449v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.23449", "identifiers": {"arxiv": "2606.23449"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.23449"}], "name": "AOHP: An Open-Source OS-Level Agent Harness for Personalized, Efficient and Secure Interaction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-22", "url": "https://arxiv.org/abs/2606.23449"}
{"aliases": [], "authors": ["Yuhang Gan", "Yiwei Yang", "Yuyi Li", "Xiangyu Gao", "Yichen Wang", "Rain Jiang", "Xiaoning Ding", "Andi Quinn", "Chen Qian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9a811375f94a751f188226d81e6dbb7257fbad25b808e5520226eadc547cfc57", "submitted_at": "2026-06-22T16:06:11Z", "updated_at": "2026-06-22T16:06:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.23521v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.23521", "identifiers": {"arxiv": "2606.23521"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.23521"}], "name": "Concordia: JIT-Compiled Persistent-Kernel Checkpointing for Fault-Tolerant LLM Inference", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-22", "url": "https://arxiv.org/abs/2606.23521"}
{"aliases": [], "authors": ["Jincheng Zhong", "Weizhi Wang", "Che Jiang", "Kai Tian", "Zhenzhao Yuan", "Junlin Yang", "Dianqiao Lei", "Kaiyan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9f59549d6987702cc24bde9be977bf8260ca8851e3a2bbb0ec665c4679765bee", "submitted_at": "2026-06-22T17:39:43Z", "updated_at": "2026-06-22T17:39:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.23654v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.23654", "identifiers": {"arxiv": "2606.23654"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.23654"}], "name": "EnterpriseClawBench: Benchmarking Agents from Real Workplace Sessions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-22", "url": "https://arxiv.org/abs/2606.23654"}
{"aliases": [], "authors": ["Marta Sumyk", "Oleksandr Kosovan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0aeb8c28b570133f575c2f336b6dab0dcbecdfaec78549115407b21722816e8a", "submitted_at": "2026-06-23T12:46:29Z", "updated_at": "2026-06-23T12:46:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.24515v1"}}, "environments": ["desktop", "macos", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.24515", "identifiers": {"arxiv": "2606.24515"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.24515"}], "name": "Reinforcement Learning for Computer-Use Agents with Autonomous Evaluation", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "The framework uses a vision-language judge’s final-screen assessment as a terminal reward for GUI reinforcement learning. It models evaluator noise and corrects the reward estimator used by PPO across desktop benchmarks.", "topics": ["reinforcement-learning", "reward-models", "evaluation", "robustness"], "updated_at": "2026-06-23", "url": "https://arxiv.org/abs/2606.24515"}
{"aliases": [], "authors": ["Jiachen Qian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a3d0341f41a8ecba57fb7ed9e4b266bb4a176639c742be2b4fc98555754f5744", "submitted_at": "2026-06-23T12:57:02Z", "updated_at": "2026-06-23T12:57:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.24525v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.24525", "identifiers": {"arxiv": "2606.24525"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.24525"}], "name": "VisCritic: Visual State Comparison as Process Reward for GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "VisCritic compares before/after screenshots through change-aware visual features and an action-conditioned critic. Weakly supervised samples from existing traces train judgments of action success, progress and error type for GUI-agent verification.", "topics": ["reward-models", "state-tracking", "data-generation", "evaluation"], "updated_at": "2026-06-23", "url": "https://arxiv.org/abs/2606.24525"}
{"aliases": [], "authors": ["Xiao Zhou", "Siyue Zhang", "Yilun Zhao", "Jinbiao Wei", "Tingyu Song", "Arman Cohan", "Chen Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77f777f05b1c169ee4ad7ae7d77e9d1fc80a0696e9a1f6eefb5028a5c3e6fe5d", "submitted_at": "2026-06-22T17:05:07Z", "updated_at": "2026-06-22T17:05:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.24551v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.24551", "identifiers": {"arxiv": "2606.24551"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.24551"}], "name": "GUI vs. CLI: Execution Bottlenecks in Screen-Only and Skill-Mediated Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "This benchmark matches desktop goals, initial states and final-state verifiers across screen-only GUI and skill-mediated CLI agents. Its experiments separate graphical execution failures from missing CLI skill coverage, including a verifier-guided skill-augmentation condition.", "topics": ["evaluation", "tool-integration", "robustness", "workflow"], "updated_at": "2026-06-22", "url": "https://arxiv.org/abs/2606.24551"}
{"aliases": [], "authors": ["Wei Zhou", "Xuanhe Zhou", "Shaokun Han", "Hongming Xu", "Guoliang Li", "Zhiyu Li", "Feiyu Xiong", "Fan Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "46d4ea352e1d67c1344c97826a92e2f7498101eabb3089d61f66b0ba2a3a7a2f", "submitted_at": "2026-06-23T16:34:55Z", "updated_at": "2026-06-23T16:34:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.24775v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.24775", "identifiers": {"arxiv": "2606.24775"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.24775"}], "name": "Are We Ready For An Agent-Native Memory System?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-23", "url": "https://arxiv.org/abs/2606.24775"}
{"aliases": [], "authors": ["Shayon Dasgupta", "Avijit Dasgupta", "C. V. Jawahar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "358de1e89a85d6e1112824adb40b53e6a0e98cc13af5da6bf189c28f5d83eabe", "submitted_at": "2026-06-23T18:44:59Z", "updated_at": "2026-06-23T18:44:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.25084v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.25084", "identifiers": {"arxiv": "2606.25084"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.25084"}], "name": "Are We There Yet? Exploring the Capabilities of MLLMs in Assistive AI Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-23", "url": "https://arxiv.org/abs/2606.25084"}
{"aliases": [], "authors": ["Aradhana Nayak", "Mussadiq Nazeer", "Wang Peng", "Feng Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c5a2d0e461b011e2d5b9dbd73012d007261d8df37bd1d687fa3bd86dc6f2246", "submitted_at": "2026-06-24T11:18:01Z", "updated_at": "2026-06-24T11:18:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.25705v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.25705", "identifiers": {"arxiv": "2606.25705"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.25705"}], "name": "GUI agent: Guided Exploration of User-Sensitive Screens", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "This short paper develops an explorer agent that expands from a demonstrated GUI task to discover requests likely to reach user-sensitive screens. The proposed exploration supports identifying situations where task execution should be handed back to the user.", "topics": ["safety-security", "human-in-the-loop", "data-generation"], "updated_at": "2026-06-24", "url": "https://arxiv.org/abs/2606.25705"}
{"aliases": [], "authors": ["Divake Kumar", "Sina Tayebati", "Devashri Naik", "Amanda Sofie Rios", "Nilesh Ahuja", "Omesh Tickoo", "Ranganath Krishnan", "Amit Ranjan Trivedi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "628d08bdbf23ae9e599bf767f0503cecedcde72dfbfcc183f9ea9f044c6f2d44", "submitted_at": "2026-06-24T12:34:28Z", "updated_at": "2026-09-29T20:55:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.25760v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.25760", "identifiers": {"arxiv": "2606.25760"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.25760"}], "name": "Uncertainty Quantification for Computer-Use Agents: A Benchmark across Vision-Language Models and GUI Grounding Datasets", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Argus benchmarks post-hoc uncertainty methods for single-step GUI grounding across open-weight and closed-source observation regimes. It tests calibration, rejection, error severity and conformal click regions, finding that uncertainty rankings transfer less reliably across model/interface regimes than across datasets for one model.", "topics": ["grounding", "evaluation", "robustness", "safety-security"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2606.25760"}
{"aliases": [], "authors": ["Chengheng Li-Chen", "Zhiqian Zhou", "Hao Chen", "Nicolas Chauvin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7a8efe4b9476d7f55ca4da17621ece653be2e33bf02be0183a339a12f5fac3ca", "submitted_at": "2026-06-24T15:35:29Z", "updated_at": "2026-06-24T15:35:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.25964v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.25964", "identifiers": {"arxiv": "2606.25964"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.25964"}], "name": "WinDOM: Self-Family Distillation for Small-Model GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "WinDOM obtains bounding-box supervision from a Windows 11 web reimplementation driven through Playwright, and studies rejection-sampling distillation for small GUI grounders. It treats the depth of the supervised cold start as a GRPO choice, comparing same-size EMA and larger same-family teachers.", "topics": ["grounding", "data-generation", "imitation-learning", "reinforcement-learning"], "updated_at": "2026-06-24", "url": "https://arxiv.org/abs/2606.25964"}
{"aliases": [], "authors": ["Changdae Oh", "Wendi Li", "Seongheon Park", "Samuel Yeh", "Tanwi Mallick", "Sharon Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "325daec9e2e305b661368680a9e9a559d479082545c9d313fb0ead159d7f995a", "submitted_at": "2026-06-24T17:54:08Z", "updated_at": "2026-06-24T17:54:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.26080v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.26080", "identifiers": {"arxiv": "2606.26080"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.26080"}], "name": "Neglected Free Lunch from Post-training: Progress Advantage for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-24", "url": "https://arxiv.org/abs/2606.26080"}
{"aliases": [], "authors": ["Tianyi Men", "Zhuoran Jin", "Pengfei Cao", "Yubo Chen", "Kang Liu", "Jun Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "42166ac3980296a53b56bc4a076634beec36ba885b3260b5d36115b5a5261cd0", "submitted_at": "2026-06-25T17:44:48Z", "updated_at": "2026-06-25T17:44:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.27330v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.27330", "identifiers": {"arxiv": "2606.27330"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.27330"}], "name": "Empowering GUI Agents via Autonomous Experience Exploration and Hindsight Experience Utilization for Task Planning", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "PEEU explores web environments and turns hindsight experience into aligned high-level planning supervision. A hierarchical task-decomposition analysis examines why atomic-skill competence alone does not ensure planning generalization across websites.", "topics": ["planning", "data-generation", "imitation-learning", "web-navigation"], "updated_at": "2026-06-25", "url": "https://arxiv.org/abs/2606.27330"}
{"aliases": [], "authors": ["Egor Bogomolov", "Yaroslav Zharov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f595d057cfccc5bdb01669f67cd8f1418a671bdb7011e0d54f63fd971813fbc", "submitted_at": "2026-06-25T08:02:32Z", "updated_at": "2026-06-25T08:02:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.27406v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.27406", "identifiers": {"arxiv": "2606.27406"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.27406"}], "name": "Towards Evaluation of Implicit Software World Models in Coding LLMs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-25", "url": "https://arxiv.org/abs/2606.27406"}
{"aliases": [], "authors": ["Minbyul Jeong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbca3da0f37ef8d6b61e7ad3498c24a4d5b111cc51021f31d59e75a44eefae2b", "submitted_at": "2026-06-25T22:51:59Z", "updated_at": "2026-06-25T22:51:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.27595v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.27595", "identifiers": {"arxiv": "2606.27595"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.27595"}], "name": "Ko-WideSearch: A Korean Breadth-Search Benchmark for Exhaustive Set Enumeration by Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-25", "url": "https://arxiv.org/abs/2606.27595"}
{"aliases": [], "authors": ["Thomas S. Paula", "Lucas S. Kupssinskü", "Rodrigo C. Barros"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aeca73d2d9f1e40e6a0bb14abf6eecd6ab803c1d0731dc9a35dc8a4eb583a505", "submitted_at": "2026-06-26T01:16:03Z", "updated_at": "2026-09-12T14:38:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.27634v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.27634", "identifiers": {"arxiv": "2606.27634"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.27634"}], "name": "Continual Learning for Sequential Personalization of Small Language Models: A Stability Monitoring Analysis", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-12", "url": "https://arxiv.org/abs/2606.27634"}
{"aliases": [], "authors": ["Yanuo Ma", "Ben Kereopa-Yorke", "Ben Schultz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a5b2051d7eff17f4a1f1ca93e29f9ab29e65c7c3e4e3aa504892a315c05672a", "submitted_at": "2026-06-26T01:23:27Z", "updated_at": "2026-06-26T01:23:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.28430v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.28430", "identifiers": {"arxiv": "2606.28430"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.28430"}], "name": "Building to the Test: Coding Agents Deliver What You Check, Not What You Requested", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-26", "url": "https://arxiv.org/abs/2606.28430"}
{"aliases": [], "authors": ["Shoufa Chen", "Luyuan Wang", "Xuan Yang", "Zhiheng Liu", "Yuren Cong", "Yuanfeng Ji", "Feiyan Zhou", "Xiaohui Zhang", "Fanny Yang", "Belinda Zeng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6164c42541650e5120adc881b19524cb3b268f08d6eb77fbe91d8c1a02823786", "submitted_at": "2026-06-26T17:59:51Z", "updated_at": "2026-06-26T17:59:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.28480v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.28480", "identifiers": {"arxiv": "2606.28480"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.28480"}], "name": "TUA-Bench: A Benchmark for General-Purpose Terminal-Use Agents", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent general computer-operation benchmark: the evaluated interface is a terminal; GUI benchmarks are a contrast, not its execution modality.", "published_at": "2026-06-26", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "TUA-Bench evaluates terminal-use agents on general digital tasks, including document editing, email, web information seeking and specialist software workflows. Tasks run in real terminals with deterministic setup and execution-based scoring.", "topics": ["evaluation", "automation"], "updated_at": "2026-06-26", "url": "https://arxiv.org/abs/2606.28480"}
{"aliases": [], "authors": ["Yuanhong Cai", "Xiaohui Nie", "Kanglin Yin", "Changhua Pei", "Yongqian Sun", "Shenglin Zhang", "Haibin Liu", "Guiyang Liu", "Xidao Wen", "Fang Situ", "Dan Pei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd03c359ad7cd8ffafe00cdcd66db2dd8a3832c035a67ca6762f156eaad67db3", "submitted_at": "2026-06-28T04:38:05Z", "updated_at": "2026-06-28T04:38:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.29193v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2606.29193", "identifiers": {"arxiv": "2606.29193"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.29193"}], "name": "A Multi-Dataset Benchmark for Evaluating LLM Agents in Microservice Failure Diagnosis", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-28", "url": "https://arxiv.org/abs/2606.29193"}
{"aliases": [], "authors": ["Sunqi Fan", "Qingle Liu", "Runqi Yin", "Meng-Hao Guo", "Shuojin Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "21fd84f29b0f5892d79714794b92b9534adb5eb459e32638163b6d062fe6f047", "submitted_at": "2026-06-28T15:11:19Z", "updated_at": "2026-06-28T15:11:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.29445v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.29445", "identifiers": {"arxiv": "2606.29445"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.29445"}], "name": "Bridging VideoQA and Video-Guided Agentic Tasks via Generalized Keyframe Extraction", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-28", "url": "https://arxiv.org/abs/2606.29445"}
{"aliases": [], "authors": ["Bojie Li", "Noah Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54de4e93f4f3e39d904b4661a193765e02c29f8ec7787264aee5616579c32210", "submitted_at": "2026-06-28T15:59:31Z", "updated_at": "2026-06-28T15:59:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.29472v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.29472", "identifiers": {"arxiv": "2606.29472"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.29472"}], "name": "Agent-Computer Observation Interfaces Enable Dynamic Computer Use", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "The Agent-Computer Observation Interface separates continuous perception from discrete computer actions using captured frames, audio transcription and persistent visual narration. DynaCU-Bench tests these components on dynamic browser tasks and static controls, with model-specific ablations showing that additional observations can also hurt.", "topics": ["evaluation", "video", "screen-understanding", "memory", "training-free"], "updated_at": "2026-06-28", "url": "https://arxiv.org/abs/2606.29472"}
{"aliases": [], "authors": ["Mengqi Yuan", "Zilong Zhou", "Xinzhuang Xiong", "Weiming Wu", "Jiayang Sun", "Jiamin Song", "Kaiqian Cui", "Bowen Wang", "Haoyuan Wu", "Yitong Li", "Dunjie Lu", "Haikong Lu", "Qi Zhen", "Xinyuan Wang", "Jiaqi Deng", "Yuhao Yang", "Cheng Chen", "Boyuan Zheng", "Alex Su", "Xiao Yu", "Hao Zou", "Saaket Agashe", "Xing Han Lu", "Manpreet Kaur", "Zhengyang Qi", "Vincent Sunn Chen", "Frederic Sala", "Dayiheng Liu", "Junyang Lin", "Zhou Yu", "Yu Su", "Siva Reddy", "Xin Eric Wang", "Peng Qi", "Tianbao Xie", "Tao Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03035fcbc6cd3e6bf40f244870fb35694ba985ff7f809da66e2a376cbf99d66b", "submitted_at": "2026-06-28T17:59:17Z", "updated_at": "2026-07-13T10:30:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.29537v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.29537", "identifiers": {"arxiv": "2606.29537"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.29537"}], "name": "OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces extended computer-use workflows with realistic artifacts, changing information, hidden state, and partial-progress and safety assessments.", "topics": ["evaluation", "long-horizon", "workflow"], "updated_at": "2026-07-13", "url": "https://arxiv.org/abs/2606.29537"}
{"aliases": [], "authors": ["Sunqi Fan", "Lingshan Chen", "Runqi Yin", "Qingle Liu", "Yongming Rao", "Meng-Hao Guo", "Shi-Min Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7beb70ac91463b77a8babaf2b077c46c5072e8f86d22f4d9afc9c1e77680f309", "submitted_at": "2026-06-29T02:16:21Z", "updated_at": "2026-06-29T02:16:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.29705v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.29705", "identifiers": {"arxiv": "2606.29705"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.29705"}], "name": "GUICrafter: Weakly-Supervised GUI Agent Leveraging Massive Unannotated Screenshots", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-06-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUICrafter first learns grounding from large collections of unannotated screenshots and webpages, then uses a smaller high-quality dataset for reinforcement-learning calibration. The curriculum targets reduced dependence on manually annotated GUI training data.", "topics": ["grounding", "data-generation", "reinforcement-learning", "screen-understanding"], "updated_at": "2026-06-29", "url": "https://arxiv.org/abs/2606.29705"}
{"aliases": [], "authors": ["Jian Zhu", "Yuzheng Zhang", "Zeyao Ma", "Bohan Zhang", "Armin Schoepf", "Daniel Woloch", "Peter Yiliu Wang", "Guangyu Robert Yang", "Samuel Jacob", "Siddharth Nagisetty", "Abhiram Chundru", "Jean Lin", "Spencer Mateega", "Jing Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b5691e4b07d0d9f49bdc3c3f5282d05b7a278f4aebe2dc341586ae275f5847c", "submitted_at": "2026-06-29T08:33:52Z", "updated_at": "2026-06-29T08:33:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.29955v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.29955", "identifiers": {"arxiv": "2606.29955"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.29955"}], "name": "SpreadsheetBench 2: Evaluating Agents on End-to-End Business Spreadsheet Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-29", "url": "https://arxiv.org/abs/2606.29955"}
{"aliases": [], "authors": ["Chen Liu", "Ling Chen", "Hanzhang Zhou", "Liangyu Chen", "Chenglin Cai", "Xin Yu", "Steven Hoi", "Yue Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50e44e08750d4152f9b5408885f8a6a016c5f789af9712e447ee147fb25adb8d", "submitted_at": "2026-06-29T10:20:39Z", "updated_at": "2026-06-29T10:20:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30084v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.30084", "identifiers": {"arxiv": "2606.30084"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.30084"}], "name": "One Forward Beats Two: InnerZoom for Accurate and Efficient GUI Grounding", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "InnerZoom preserves target-region cues from intermediate decoder layers and reinjects a compact evidence state during coordinate decoding. Its single-forward GUI grounder is compared with crop-and-rerun zooming and matched supervised/RL baselines for accuracy and inference cost.", "topics": ["grounding", "efficiency", "screen-understanding", "evaluation"], "updated_at": "2026-06-29", "url": "https://arxiv.org/abs/2606.30084"}
{"aliases": [], "authors": ["Iliana Fayolle", "Sihem Bouhenniche", "Samuel Pélissier", "Pierre Laperdrix", "Clémentine Maurice", "Walter Rudametkin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "770c901e3490b2a43b02296136825e8e6bd035fb7494ea74b3a9af74be1afed1", "submitted_at": "2026-06-29T10:51:13Z", "updated_at": "2026-06-29T10:51:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30119v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.30119", "identifiers": {"arxiv": "2606.30119"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.30119"}], "name": "On the Internet, Nobody Knows You're an LLM Bot: Unmasking Web Agents with Multi-Layer Fingerprinting", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates web agents against anti-bot mechanisms on instrumented honeysites and analyzes network, HTTP and browser fingerprints. The study compares resistance to blocking with the ability to distinguish agents, humans and agent implementations.", "topics": ["safety-security", "observability", "evaluation"], "updated_at": "2026-06-29", "url": "https://arxiv.org/abs/2606.30119"}
{"aliases": [], "authors": ["Lei Bai", "Zongsheng Cao", "Yang Chen", "Zhiyao Cui", "Shangheng Du", "Yue Fan", "Shiyang Feng", "Zijie Guo", "Haonan He", "Liang He", "Xiaohan He", "Shuyue Hu", "Yusong Hu", "Songtao Huang", "Yichen Jiang", "Hao Li", "Xin Li", "Dahua Lin", "Weihao Lin", "Fenghua Ling", "Dongrui Liu", "Zhuo Liu", "Wenjie Lou", "Runmin Ma", "Chunjiang Mu", "Haoyang Peng", "Tianshuo Peng", "Jinxin Shi", "Luohe Shi", "Boyuan Sun", "Zelin Tan", "Shengji Tang", "Yan Teng", "Qianyi Wang", "Xiaosong Wang", "Yiming Wu", "Yi Xie", "Xiangchao Yan", "Jingqi Ye", "Peng Ye", "Fangchen Yu", "Jiakang Yuan", "Bihao Zhan", "Bo Zhang", "Chen Zhang", "Shufei Zhang", "Shuaiyu Zhang", "Wenlong Zhang", "Yiqun Zhang", "Junpeng Zhao", "Zhijie Zhong", "Bowen Zhou", "Yuhao Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae4eacb6be976230dab57b6e2acd8adeec8461e0a6cb2ee43ef6c33249f74b89", "submitted_at": "2026-06-29T17:50:54Z", "updated_at": "2026-07-13T13:01:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.30616v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2606.30616", "identifiers": {"arxiv": "2606.30616"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.30616"}], "name": "Scaling the Horizon, Not the Parameters: Reaching Trillion-Parameter Performance with a 35B Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-13", "url": "https://arxiv.org/abs/2606.30616"}
{"aliases": [], "authors": ["Yogeswar Reddy Thota"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f3f2ae1438986c11146fec508dd61fd3e848ddcbf830f86a7dfa759280d730a", "submitted_at": "2026-06-29T03:19:40Z", "updated_at": "2026-06-29T03:19:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30697v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.30697", "identifiers": {"arxiv": "2606.30697"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.30697"}], "name": "LUMOS: A Semantic Operating-System Layer for Accessibility-Grounded AI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "LUMOS converts native accessibility metadata and browser UI structure into semantic interface blueprints with stable element identifiers and available actions. It combines live pointer-to-element grounding with an observe-act loop constrained to visible-UI primitives.", "topics": ["grounding", "screen-understanding", "tool-integration", "automation"], "updated_at": "2026-06-29", "url": "https://arxiv.org/abs/2606.30697"}
{"aliases": [], "authors": ["Apurva Gandhi", "Vishwas Suryanarayanan", "Raja Hasnain Anwar", "Firoz Shaik", "Shubhang Desai", "Thong Q. Nguyen", "Muhammad Taqi Raza", "Vishal Chowdhary", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47d2057e8cf064eac0e50c621a6fb2cccacc2aa57f92b9a552753d38c224c1ad", "submitted_at": "2026-06-30T05:26:51Z", "updated_at": "2026-06-30T05:26:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.31154v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.31154", "identifiers": {"arxiv": "2606.31154"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.31154"}], "name": "PPT-Eval: A Benchmark for Computer-Use Agents on PowerPoint Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-30", "url": "https://arxiv.org/abs/2606.31154"}
{"aliases": [], "authors": ["Xueqiao Sun", "Xiaohan Wang", "Ludwig Schmidt", "Serena Yeung-Levy", "Yuhui Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8911abdc0c9c481d82abb213d368c16fba9fee909c354e0bc6272e14c59abfb7", "submitted_at": "2026-06-30T07:44:37Z", "updated_at": "2026-06-30T07:44:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.31270v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.31270", "identifiers": {"arxiv": "2606.31270"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.31270"}], "name": "Learning from Failure: Inference-Time Self-Improvement for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-06-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses failed computer-use trajectories to diagnose failure modes and generate inference-time agent code patches, with light human verification. An OSWorld study evaluates this failure-driven improvement loop without additional model training.", "topics": ["self-reflection", "robustness", "training-free", "evaluation"], "updated_at": "2026-06-30", "url": "https://arxiv.org/abs/2606.31270"}
{"aliases": [], "authors": ["Wanxia Cao", "Chengzhen Duan", "Pei Fu", "Pengzhi Gao", "Niu Lian", "Fazhan Liu", "Hui Liu", "Heng Qu", "Qinzhuo Wu", "Zhehao Yu", "Tongbo Chen", "Shiqi Cui", "Anan Du", "Shukai Jia", "Yuanfa Li", "Wei Liu", "Yike Liu", "Wenchao Lu", "Zhenbo Luo", "Haoyuan Sun", "Jiatong Sun", "Cheng Tan", "Yajie Wang", "Changqiao Wu", "Tao Xiong", "Jiahui Yang", "Yuxuan Yuan", "Ruoceng Zhang", "Shaojie Zhang", "Jian Zhu", "Jian Luan", "Cong Zou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "974d471d25373502914bf0d146d5d4a54a5751281c04921a9d331d83bce64b33", "submitted_at": "2026-06-30T09:36:35Z", "updated_at": "2026-07-01T03:59:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.31410v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "id": "arxiv:2606.31410", "identifiers": {"arxiv": "2606.31410"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.31410"}], "name": "Xiaomi-GUI-0 Technical Report", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a mobile GUI agent trained through real-device interaction, corrected failure trajectories, supervised learning, and reinforcement learning.", "topics": ["mobile-navigation", "reinforcement-learning", "robustness"], "updated_at": "2026-07-01", "url": "https://arxiv.org/abs/2606.31410"}
{"aliases": [], "authors": ["Chen Liu", "Ling Chen", "Hanzhang Zhou", "Xu Zhang", "Quyu Kong", "Panrong Tong", "Wenhao Wang", "Xin Yu", "Steven Hoi", "Yue Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6765a1dc51ee1141d09d23dcff091495333db731d02c7067d833e14cce713925", "submitted_at": "2026-06-30T13:01:19Z", "updated_at": "2026-07-02T06:42:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.31612v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.31612", "identifiers": {"arxiv": "2606.31612"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2606.31612"}], "name": "What Memory Do GUI Agents Really Need? From Passive Records to Active Task-Driving States", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-02", "url": "https://arxiv.org/abs/2606.31612"}
{"aliases": [], "authors": ["Meng Chen", "Anya Ji", "Tsung-Han Wu", "Tobias Maringgele", "David M. Chan", "Alane Suhr", "Amy Pavel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0a3e641e0a71051f5fbd20f57af3db06f2b65c61025e78adcdd122f773d034b", "submitted_at": "2026-06-30T17:18:47Z", "updated_at": "2026-09-01T04:17:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.31980v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2606.31980", "identifiers": {"arxiv": "2606.31980"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.31980"}], "name": "DigitalCoach: Communication and Grounding Gaps in Human and Agentic Computer Use Coaching", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-06-30", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "DigitalCoach records expert–novice software coaching conversations aligned with screen and input events. Automated and interactive evaluations examine how model coaches differ from humans in explanation, error diagnosis and visual grounding when helping a learner operate software.", "topics": ["human-in-the-loop", "video", "evaluation", "grounding"], "updated_at": "2026-09-01", "url": "https://arxiv.org/abs/2606.31980"}
{"aliases": [], "authors": ["Zidong Zhang", "Zhentao Xie", "Wenrui Diao", "Jianliang Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cd4d31e477d5915626d471682e52e5cd6083d9b0c67196f6de5de2a29c6a3058", "submitted_at": "2026-07-01T02:17:19Z", "updated_at": "2026-07-14T03:01:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.00333v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2607.00333", "identifiers": {"arxiv": "2607.00333"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.00333"}], "name": "(A)I Sees What You Don't: Exploiting New Attack Surfaces in Third-Party Mobile Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes cross-app privilege escalation and visual attack surfaces introduced by third-party VLM-powered mobile agents.", "topics": ["safety-security", "mobile-navigation"], "updated_at": "2026-07-14", "url": "https://arxiv.org/abs/2607.00333"}
{"aliases": [], "authors": ["Yujie Zheng", "Zikang Liu", "Xin Zhao", "Ji-Rong Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89f885061b8e590b05ff3c2138eb0f9c7a31aac038a04d330d435dc6f228f25d", "submitted_at": "2026-07-01T06:37:21Z", "updated_at": "2026-07-01T06:37:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.00502v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.00502", "identifiers": {"arxiv": "2607.00502"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.00502"}], "name": "A Task-State Representation for Long-Horizon Mobile GUI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-01", "url": "https://arxiv.org/abs/2607.00502"}
{"aliases": [], "authors": ["Xubin Hao", "Hongjin Meng", "Xin Yin", "Jiawei Zhu", "Chenpeng Cao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9241d95c69c1f521a64932bddf77679cf8bd3547f2958ab2a24189789e3716f8", "submitted_at": "2026-07-01T09:41:54Z", "updated_at": "2026-07-01T09:41:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.00692v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2607.00692", "identifiers": {"arxiv": "2607.00692"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.00692"}], "name": "Self-GC: Self-Governing Context for Long-Horizon LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-01", "url": "https://arxiv.org/abs/2607.00692"}
{"aliases": [], "authors": ["Song-Lin Lv", "Weiming Wu", "Rui Zhu", "Zi-Jian Cheng", "Lan-Zhe Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "240b39387fd314d3feea1c193ea3a1eca3e88a875ad15a4fd8de9f6e3d0998d1", "submitted_at": "2026-07-01T15:40:25Z", "updated_at": "2026-07-01T15:40:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.01084v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2607.01084", "identifiers": {"arxiv": "2607.01084"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.01084"}], "name": "Can Agents Generalize to the Open World? Unveiling the Fragility of Static Training in Tool Use", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-01", "url": "https://arxiv.org/abs/2607.01084"}
{"aliases": [], "authors": ["Zhaoyan Sun", "Shan Zhong", "Daizhou Wen", "Jiaxing Han", "Guoliang Li", "Ying Yan", "Peng Zhang", "Yu Su", "Xiang Qi", "Baolin Sun", "Chengyuan Yang", "Tao Fang", "Huaiyu Ruan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6c3c0176c80dd7d2b0f4af439d326b82a4b61e7eb308df11c1f895eb089eb4a1", "submitted_at": "2026-07-02T03:18:59Z", "updated_at": "2026-07-02T03:18:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.01647v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2607.01647", "identifiers": {"arxiv": "2607.01647"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.01647"}], "name": "AgenticDataBench: A Comprehensive Benchmark for Data Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-02", "url": "https://arxiv.org/abs/2607.01647"}
{"aliases": [], "authors": ["Yueqi Song", "Lintang Sutawika", "Jiarui Liu", "Lindia Tjuatja", "Jiayi Geng", "Yunze Xiao", "Daniel Lee", "Aditya Bharat Soni", "Vincent Lo", "Xiang Yue", "Graham Neubig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b257174ef7d246cf0bc8567b911eb0a1e3a73fc49d6785492c3c4baa3fefd1e5", "submitted_at": "2026-07-02T10:59:03Z", "updated_at": "2026-07-06T17:18:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.02032v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2607.02032", "identifiers": {"arxiv": "2607.02032"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.02032"}], "name": "PACE: A Proxy for Agentic Capability Evaluation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-06", "url": "https://arxiv.org/abs/2607.02032"}
{"aliases": [], "authors": ["Emmanuel George", "Christopher Keefe", "Peter Pak", "Amir Barati Farimani"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f79130b0f8293bf2a7b4dc23f9faf17e2d87b7f55ae424dbdfb2b3103aba9d5", "submitted_at": "2026-07-02T17:19:41Z", "updated_at": "2026-07-07T04:44:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.02448v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2607.02448", "identifiers": {"arxiv": "2607.02448"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.02448"}], "name": "AgentsCAD: Automated Design for Manufacturing of FDM Parts via Multi-Agent LLM Reasoning and Geometric Feature Recognition", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-07", "url": "https://arxiv.org/abs/2607.02448"}
{"aliases": [], "authors": ["Liyan Tang", "Fangcong Yin", "Greg Durrett"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a8b752178fe6e4c6ed5f8e33f98b4f485282e304d6ce134c037ea3927db6e4a", "submitted_at": "2026-07-02T17:53:15Z", "updated_at": "2026-07-02T17:53:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.02490v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2607.02490", "identifiers": {"arxiv": "2607.02490"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.02490"}], "name": "Visually Grounded Self-Reflection for Vision-Language Models via Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-02", "url": "https://arxiv.org/abs/2607.02490"}
{"aliases": [], "authors": ["Vishvesh Bhat", "Jay Vaghasiya", "Muhammad Ahmed Mohsin", "Asad Aali"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6ec2c2f6ed3e81522c0414f8c32b2ea8fd4c67b826bc062402b1d48658fce0a", "submitted_at": "2026-06-30T21:10:42Z", "updated_at": "2026-06-30T21:10:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.02577v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2607.02577", "identifiers": {"arxiv": "2607.02577"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.02577"}], "name": "Benchmarking the Benchmarks: A Validity Audit of Tool-Calling Evaluation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-06-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-06-30", "url": "https://arxiv.org/abs/2607.02577"}
{"aliases": [], "authors": ["Haorui Chen", "Yuancheng Zhu", "Yitong Zhang", "Jia Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "feed638e22f96d265496a0a2487d2b25e4dddeed319a0dab49db08a0f0ad0daf", "submitted_at": "2026-07-03T03:17:52Z", "updated_at": "2026-07-03T03:17:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.02911v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "id": "arxiv:2607.02911", "identifiers": {"arxiv": "2607.02911"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.02911"}], "name": "CoACT: Action-Preserving Observation Compression for Coding Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-03", "url": "https://arxiv.org/abs/2607.02911"}
{"aliases": [], "authors": ["Shu Yang", "Junchao Wu", "Derek F. Wong", "Di Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "02f73c44472973db90cc9d715360e9508a2241b0d52c8a5fd9eabf4ad09c8e15", "submitted_at": "2026-07-04T06:27:19Z", "updated_at": "2026-07-04T06:27:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.03726v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.03726", "identifiers": {"arxiv": "2607.03726"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.03726"}], "name": "SelfMem: Self-Optimizing Memory for AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-04", "url": "https://arxiv.org/abs/2607.03726"}
{"aliases": [], "authors": ["Guijia Zhang", "Yuxun Chen", "Yuheng Qi", "Harry Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7fdae5d6e5ea17639fe0e2b8e2d655df7201a438d84acdeb3e0d02a78051ae2", "submitted_at": "2026-07-05T14:33:29Z", "updated_at": "2026-07-18T16:56:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.04334v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.04334", "identifiers": {"arxiv": "2607.04334"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.04334"}], "name": "Do GUI Agents Believe Their Eyes? Diagnosing State-Belief Reliance on Pixels versus Structure", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Defines visual state reliance and probes whether GUI agents resolve screenshot/structure conflicts using pixels, serialized interface structure or prior beliefs. Paired interventions and live multi-step tests connect misattributed state beliefs to failures and compare consistency-based mitigations.", "topics": ["screen-understanding", "state-tracking", "evaluation", "robustness", "safety-security"], "updated_at": "2026-07-18", "url": "https://arxiv.org/abs/2607.04334"}
{"aliases": [], "authors": ["Niu Lian", "Tongbo Chen", "Zhehao Yu", "Chengzhen Duan", "Fazhan Liu", "Hui Liu", "Pei Fu", "Jian Luan", "Heng Qu", "Shu-Tao Xia", "Jinpeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "32cb3020ccf25c31b805dc44218f5f9496cbc310e1df2906d8b15e5b60c31278", "submitted_at": "2026-07-05T17:37:12Z", "updated_at": "2026-08-10T13:11:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.04425v2"}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.04425", "identifiers": {"arxiv": "2607.04425"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.04425"}], "name": "UI-MOPD: Multi-Platform On-Policy Distillation for Unified GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Uni-GUI, a collection of executable desktop and mobile trajectories, and UI-MOPD, which trains a shared policy from its own rollouts using platform-specific teachers. The distillation procedure routes each visited state to the appropriate teacher to preserve distinct interaction conventions.", "topics": ["data-generation", "imitation-learning", "action-prediction"], "updated_at": "2026-08-10", "url": "https://arxiv.org/abs/2607.04425"}
{"aliases": [], "authors": ["Woohyuk Choi", "Juhee Kim", "Taehyun Kang", "Jihyeon Jeong", "Luyi Xing", "Byoungyoung Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5fb30b8be2c6e452471e6398720291ec59a2c5ef96e64baca0735fb01b7a12e", "submitted_at": "2026-07-06T14:07:49Z", "updated_at": "2026-07-06T14:07:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.05120v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.05120", "identifiers": {"arxiv": "2607.05120"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.05120"}], "name": "Agent Data Injection Attacks are Realistic Threats to AI Agents", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-07-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study examines agent data injection, where untrusted content impersonates trusted metadata or tool-context data. It reports arbitrary-click vulnerabilities in web agents alongside separate coding-agent attacks, highlighting trust-boundary failures beyond instruction injection.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-07-06", "url": "https://arxiv.org/abs/2607.05120"}
{"aliases": [], "authors": ["Kristina Nikolić", "Egor Zverev", "Javier Rando", "Matthew Jagielski", "Edoardo Debenedetti", "Florian Tramèr"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97d6c625b6ebc08cec7471df543812a91c4252f50d8d35f8d51fbef802dcb5a2", "submitted_at": "2026-07-06T16:22:37Z", "updated_at": "2026-07-06T16:22:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.05277v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.05277", "identifiers": {"arxiv": "2607.05277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.05277"}], "name": "Untrusted Content Masking for Web Agents with Security Guarantees", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Untrusted Content Masking uses DOM structure to redact untrusted webpage regions before agent observation. A sandboxed interaction interface separates privileges so browser actions can proceed under the proposed trust-boundary model without exposing protected content to the planner.", "topics": ["safety-security", "sandbox", "web-navigation"], "updated_at": "2026-07-06", "url": "https://arxiv.org/abs/2607.05277"}
{"aliases": [], "authors": ["Hairui Zhu", "Yiying Yang", "Tengjin Weng", "Ziyu Lu", "Xiao Yao", "Xiaoyang Ye", "Lin Ma", "Wenhao Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c7b0105af9699ebe985c88b4aef0891df1597453d870809b261fd6434ba03756", "submitted_at": "2026-07-06T04:57:18Z", "updated_at": "2026-07-06T04:57:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.05465v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.05465", "identifiers": {"arxiv": "2607.05465"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.05465"}], "name": "CanvasAgent: Enabling Complex Image Creation and Editing via Visual Tool Orchestration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-06", "url": "https://arxiv.org/abs/2607.05465"}
{"aliases": [], "authors": ["Mohammadreza Rashidi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ddac32c59f6efbecfa9b02407b2fc9823ffc2bb122e5af241b0decb3b7e06031", "submitted_at": "2026-07-07T01:56:18Z", "updated_at": "2026-07-07T01:56:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.05743v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.05743", "identifiers": {"arxiv": "2607.05743"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.05743"}], "name": "The Balkanization of Execution-Security Research for AI Coding Agents: Isolation, Access Control, and Time-of-Check-to-Time-of-Use Vulnerabilities", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-07", "url": "https://arxiv.org/abs/2607.05743"}
{"aliases": [], "authors": ["Hongliang Li", "Yijin Liu", "Zhiwei Zhang", "Zihe Liu", "Xinyue Lou", "Jinan Xu", "Fandong Meng", "Kaiyu Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b31ef8e43f5ef5da8224e655ad2bb2e916216f11a1d89e4afe298fa2a1c57a97", "submitted_at": "2026-07-07T08:50:09Z", "updated_at": "2026-08-17T05:33:41Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.06008v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.06008", "identifiers": {"arxiv": "2607.06008"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.06008"}], "name": "PolyWorkBench: Benchmarking LLM Agents for Cross-Lingual Long-Horizon Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-17", "url": "https://arxiv.org/abs/2607.06008"}
{"aliases": [], "authors": ["Wei Dong", "Tianyu Fu", "Zhe Yu", "Hanning Wang", "Anyang Su", "Zhizhou Fang", "Yuyang Chen", "Shuo Wang", "Minghui Wu", "Ping Jiang", "Zhen Lei", "Chenxu Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc6d846f43b9992388e94f3bcfdfc9a25d09b317fea62336aa22951c8fce9f8b", "submitted_at": "2026-07-07T10:27:31Z", "updated_at": "2026-07-07T10:27:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.06118v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.06118", "identifiers": {"arxiv": "2607.06118"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.06118"}], "name": "WebRetriever: A Large-Scale Comprehensive Benchmark for Efficient Web Agent Evaluation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebRetriever evaluates navigation, knowledge-assisted interaction and end-to-end information-extraction workflows under separate protocols. Its NavEval judge uses richer interaction context to diagnose cases where navigation success does not translate into useful task completion.", "topics": ["evaluation", "web-navigation", "observability"], "updated_at": "2026-07-07", "url": "https://arxiv.org/abs/2607.06118"}
{"aliases": [], "authors": ["Grace Man Chen", "Litao Guo", "Yifan Wu", "Yiyu Chen", "Yenchi Tseng", "Sicheng Liu", "Yuyu Luo", "Ying-Cong Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff483463f0eda2e33526548ee91caeeccf719d88805ed90f0481876c71be150a", "submitted_at": "2026-07-07T14:08:55Z", "updated_at": "2026-10-04T02:42:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.06306v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.06306", "identifiers": {"arxiv": "2607.06306"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.06306"}], "name": "UI2App: Benchmarking Visual Interaction Inference in Executable Web Application Generation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-04", "url": "https://arxiv.org/abs/2607.06306"}
{"aliases": [], "authors": ["George Torres", "Sharad Shrestha", "Satyajayant Misra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89efe79458032881cf6b6975011a62efe1ae90b246c1cf8494aaac4f94e2fe58", "submitted_at": "2026-07-06T19:30:53Z", "updated_at": "2026-07-06T19:30:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.06595v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.06595", "identifiers": {"arxiv": "2607.06595"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.06595"}], "name": "When Agents Remember Too Much: Memory Poisoning Attacks on Large Language Model Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-06", "url": "https://arxiv.org/abs/2607.06595"}
{"aliases": [], "authors": ["Yingyun Cui", "Yi Xie", "Piaohong Wang", "Jiawei Ma", "Bo Liu", "Liangliang Cao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7729f8915c4f4bd2486b71ab0b10b6bffab4624456e57f5ccd3388485e09431", "submitted_at": "2026-07-08T08:32:18Z", "updated_at": "2026-07-08T08:32:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.07744v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.07744", "identifiers": {"arxiv": "2607.07744"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.07744"}], "name": "PERFOPT-Bench: Evaluating Coding Agents on Software Performance Optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-08", "url": "https://arxiv.org/abs/2607.07744"}
{"aliases": [], "authors": ["Xinyu Geng", "Xuanhua He", "Sixiang Chen", "Yanjing Xiao", "Fan Zhang", "Shijue Huang", "Haitao Mi", "Zhenwen Liang", "Tianqing Fang", "Yi R. Fung"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff02f4f78756087698329c631c0d3d186bebc418207c714d427c96985bdee356", "submitted_at": "2026-07-08T18:03:41Z", "updated_at": "2026-07-13T12:09:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.07820v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.07820", "identifiers": {"arxiv": "2607.07820"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.07820"}], "name": "DeepSearch-World: Self-Distillation for Deep Search Agents in a Verifiable Environment", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-13", "url": "https://arxiv.org/abs/2607.07820"}
{"aliases": [], "authors": ["Corban Villa", "Alp Eren Ozdarendeli", "Sijun Tan", "Raluca Ada Popa"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f2e44014824a9bf9590885c227482da71a305f461de88b5b6ff6adb01ad42d7f", "submitted_at": "2026-07-09T06:37:52Z", "updated_at": "2026-07-09T06:37:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.08147v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.08147", "identifiers": {"arxiv": "2607.08147"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.08147"}], "name": "Prismata: Confining Cross-Site Prompt Injection in Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-09", "url": "https://arxiv.org/abs/2607.08147"}
{"aliases": [], "authors": ["Mengchen Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0034a931bd08231590b57e8bd58c27f8a6d6daf03b151708596b86e3a34d583f", "submitted_at": "2026-07-09T08:56:30Z", "updated_at": "2026-10-02T08:56:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.08252v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.08252", "identifiers": {"arxiv": "2607.08252"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.08252"}], "name": "A Multi-Timescale Recursive Self-Improvement Engine for Open-Ended Persona Growth", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-02", "url": "https://arxiv.org/abs/2607.08252"}
{"aliases": [], "authors": ["Zongxia Li", "Zhongzhi Li", "Yucheng Shi", "Ruhan Wang", "Junyao Yang", "Zhichao Liu", "Xiyang Wu", "Anhao Li", "Chengsong Huang", "Ergan Shang", "Yue Yu", "Ninghao Liu", "Lichao Sun", "Haotao Mi", "Leowei Liang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c19d420efb8756981a464b0c6ef6d1adb01ddc50a6bf6594fc17a94d30e83cc8", "submitted_at": "2026-07-09T21:56:37Z", "updated_at": "2026-10-09T17:16:07Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.08964v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.08964", "identifiers": {"arxiv": "2607.08964"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.08964"}], "name": "Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2607.08964"}
{"aliases": [], "authors": ["Mianqiu Huang", "Taofeng Xue", "Chong Peng", "Jinrui Ding", "Jie Yang", "Sicheng Fan", "Jiale Hong", "Yufei Gao", "Xiaocheng Zhang", "Linsen Guo", "Xin Yang", "Dengchang Zhao", "Yuchen Xie", "Peng Pei", "Xunliang Xie", "Xipeng Qiu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53c254c53a834d4f28bb4b65317121f5c46f5dd6985d03a30f1cc1f0fa78707d", "submitted_at": "2026-07-07T16:36:25Z", "updated_at": "2026-09-04T15:47:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.09773v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.09773", "identifiers": {"arxiv": "2607.09773"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.09773"}], "name": "EvoCUA-1.5: Online Reinforcement Learning for Multi-turn Computer-Use Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "EvoCUA-1.5 trains multi-turn desktop agents through online reinforcement learning in executable sandboxes. It combines step-level policy optimization, task filtering and calibration, an adaptive curriculum, and asynchronous rollout handling with staleness control.", "topics": ["reinforcement-learning", "long-horizon", "sandbox", "data-generation"], "updated_at": "2026-09-04", "url": "https://arxiv.org/abs/2607.09773"}
{"aliases": [], "authors": ["Chengguang Gan", "Hanjun Wei", "Yunhao Liang", "Zhixi Cai", "Qinghao Zhang", "Shiwen Ni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37cfa436d182ed8fbc8823213b9173fbfcf9fb421e8d2ae89a093a5393c374ba", "submitted_at": "2026-07-11T02:08:46Z", "updated_at": "2026-07-16T03:28:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.10079v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.10079", "identifiers": {"arxiv": "2607.10079"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.10079"}], "name": "MAG: A Web-Agent Benchmark and Harness for Multimodal Action and Guide Generation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "MAG jointly evaluates browser task actions and user-facing guides grounded in screenshots through element marks or pixel coordinates. Its harness covers annotation, live evaluation and joint metrics, with an expert-trajectory-augmented GRPO training study.", "topics": ["evaluation", "grounding", "reinforcement-learning", "action-prediction"], "updated_at": "2026-07-16", "url": "https://arxiv.org/abs/2607.10079"}
{"aliases": [], "authors": ["Kefan Song", "Yanjun Qi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "96ab0ce100fab079bb552730b2ec45ff1d4d7f36050b38965c2dfc3df0a1121c", "submitted_at": "2026-07-11T19:39:47Z", "updated_at": "2026-07-11T19:39:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.10455v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.10455", "identifiers": {"arxiv": "2607.10455"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.10455"}], "name": "ANCHOR: Automated Alignment Auditing for CLI Agents on Real-World Harm", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-11", "url": "https://arxiv.org/abs/2607.10455"}
{"aliases": [], "authors": ["Yixiong Chen", "Xinyi Bai", "Alan Yuille"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0473ced7881ed2007aa49c2767860c33719d5bf1c798dfe4771a4061b13d655e", "submitted_at": "2026-07-12T07:05:31Z", "updated_at": "2026-07-12T07:05:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.10608v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.10608", "identifiers": {"arxiv": "2607.10608"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.10608"}], "name": "The Compliance Trap: Diagnosing How AI Agents Consume Conflicting Memory", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-07-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "The Entry–Propagation–Recovery framework traces where conflicting memory first changes an agent action, how the effect propagates and whether recovery occurs. WebArena and the controlled MemTrapBench separate memory-consumption failures across these stages.", "topics": ["memory", "safety-security", "robustness", "evaluation", "long-horizon"], "updated_at": "2026-07-12", "url": "https://arxiv.org/abs/2607.10608"}
{"aliases": [], "authors": ["Bowen Lv", "Xiao Liu", "Yanyu Ren", "Hanyu Lai", "Bohao Jing", "Hanchen Zhang", "Yanxiao Zhao", "Shuntian Yao", "Jie Tang", "Yuxiao Dong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "554844f23cddbdaadc3c52e93d07d06c756faadbd0136ffef0984b25423f84f8", "submitted_at": "2026-07-13T07:32:35Z", "updated_at": "2026-07-13T07:32:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.11185v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.11185", "identifiers": {"arxiv": "2607.11185"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.11185"}], "name": "SCALECUA: Scaling Computer Use Agents with Verifiable Task Synthesis and Efficient Online RL", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Scales computer-agent reinforcement learning through verifiable task generation, capability-aware rollout allocation, and segmented visual context.", "topics": ["data-generation", "reinforcement-learning", "efficiency"], "updated_at": "2026-07-13", "url": "https://arxiv.org/abs/2607.11185"}
{"aliases": [], "authors": ["Wenyi Wu", "Sibo Zhu", "Kun Zhou", "Aayush Salvi", "Zixuan Song", "Biwei Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40daa1b180a0c2eaba7b907a0f61b8dd8819cb3033b9da118162330d97cfecf7", "submitted_at": "2026-07-13T10:48:25Z", "updated_at": "2026-07-13T10:48:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.11388v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.11388", "identifiers": {"arxiv": "2607.11388"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.11388"}], "name": "StructAgent: Harness Long-horizon Digital Agents with Unified Causal Structure", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "StructAgent maintains compact task state and regulates long-horizon execution through verified transitions and progress checkpoints. Its workflow supports evidence-based completion and targeted recovery, with desktop computer-use evaluation and a separate Minecraft extension.", "topics": ["state-tracking", "planning", "long-horizon", "robustness"], "updated_at": "2026-07-13", "url": "https://arxiv.org/abs/2607.11388"}
{"aliases": [], "authors": ["Said Elnaffar", "Farzad Rashidi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "78eb2b2b4aa393934f615fdd25d5c3887e117fe437e1f6ea33fd0a2ca8e7fa01", "submitted_at": "2026-07-13T18:16:58Z", "updated_at": "2026-07-13T18:16:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.12056v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.12056", "identifiers": {"arxiv": "2607.12056"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.12056"}], "name": "Designing Agent-Ready Websites for AI Web Agents: A Framework for Machine Readability, Actionability, and Decision Reliability", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-13", "url": "https://arxiv.org/abs/2607.12056"}
{"aliases": [], "authors": ["Huihao Jing", "Wenbin Hu", "Shaojin Chen", "Haochen Shi", "Sirui Zhang", "Hanyu Yang", "Changxuan Fan", "Zhongwei Xie", "Hongyu Luo", "Wun Yu Chan", "Wei Fan", "Haoran Li", "Yangqiu Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "13679b9680f66a50eb398cd04ab6696cdcbca4affdd62d6c6b6ab0b0be17d77e", "submitted_at": "2026-07-14T06:25:35Z", "updated_at": "2026-09-02T10:25:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.12406v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.12406", "identifiers": {"arxiv": "2607.12406"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.12406"}], "name": "Isolation as a First-Class Principle for LLM-Agent System Safety: Concepts, Taxonomy, Challenges and Future Directions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-02", "url": "https://arxiv.org/abs/2607.12406"}
{"aliases": [], "authors": ["Yunxin Li", "Jinchao Li", "Shibo Su", "Zhenran Xu", "Chenrui Zhao", "Tongshu Bian", "Xiaoman Liang", "Meishan Zhang", "Baotian Hu", "Min Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "281de2bb6e45a58efd047735accc2049deab0f8815931c86abbd38b86b6d23bf", "submitted_at": "2026-07-14T11:04:01Z", "updated_at": "2026-07-15T04:41:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.12625v2"}}, "environments": ["desktop", "android", "ios", "harmonyos", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.12625", "identifiers": {"arxiv": "2607.12625"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.12625"}], "name": "KnowAct-GUIClaw: Know Deeply, Act Perfectly, Personal GUI Assistant with Self-Evolving Memory and Skill", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "KnowAct-GUIClaw combines a host planner with pluggable GUI subagents, attributable experience memory and an evolving skill library. It uses user profiles and feedback for task decomposition and evaluates transfer across Android, iOS, HarmonyOS and Windows.", "topics": ["memory", "planning", "interaction-history", "mobile-navigation"], "updated_at": "2026-07-15", "url": "https://arxiv.org/abs/2607.12625"}
{"aliases": [], "authors": ["Chengguang Gan", "Zhixi Cai", "Yunhao Liang", "Hanjun Wei", "Shiwen Ni", "Qinghao Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e96ea0cf71ff705c064db1bcd1d2cbf8b8cd712bb4e8030931cafbb31ff29b10", "submitted_at": "2026-07-14T11:17:35Z", "updated_at": "2026-07-14T11:17:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.12640v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.12640", "identifiers": {"arxiv": "2607.12640"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.12640"}], "name": "A Learning-Rate-Gated Failure of GRPO in a Small Language and Vision-Language Model Web Agent: A Controlled Null and Its Mechanism", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies a controlled setting where GRPO fails to improve a strong small web agent and can reduce text-track performance. Matched reachable-reward tasks and component grafting distinguish limited optimization headroom from learning-rate-dependent failure mechanisms.", "topics": ["reinforcement-learning", "evaluation", "robustness"], "updated_at": "2026-07-14", "url": "https://arxiv.org/abs/2607.12640"}
{"aliases": [], "authors": ["Yupeng Zheng", "Kai Zou", "Bin Liu", "Nenghai Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "642b1eed13ec86dd4a7e0354ea9b620d8da77c83322eff01aca3d6b3c95d918f", "submitted_at": "2026-07-14T13:21:36Z", "updated_at": "2026-08-07T13:24:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.12756v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.12756", "identifiers": {"arxiv": "2607.12756"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.12756"}], "name": "VisCo: Leveraging Large Language Models as Intrinsic Encoders for Visual Token Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-07", "url": "https://arxiv.org/abs/2607.12756"}
{"aliases": [], "authors": ["Hongru Cai", "Yongqi Li", "Ran Wei", "Wenjie Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "415b0502976dba375f39bded0ad3d5371dc83fd2248cea395063b776b3b129be", "submitted_at": "2026-07-14T17:58:57Z", "updated_at": "2026-09-03T15:20:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.13027v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.13027", "identifiers": {"arxiv": "2607.13027"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.13027"}], "name": "PalmClaw: A Native On-Device Agent Framework for Mobile Phones", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-03", "url": "https://arxiv.org/abs/2607.13027"}
{"aliases": [], "authors": ["Ananda Dhakal", "Krish Neupane", "Aarjan Chaudhary"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "292a5b1a1ccb6ff4111d9250323fdfa50a8249fa11ca1f9cc5f40bf482bbd001", "submitted_at": "2026-07-13T13:31:46Z", "updated_at": "2026-07-13T13:31:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.13085v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.13085", "identifiers": {"arxiv": "2607.13085"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.13085"}], "name": "Baselines Before Architecture: Evaluating Coding Agents for Autonomous Penetration Testing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-13", "url": "https://arxiv.org/abs/2607.13085"}
{"aliases": [], "authors": ["Richmond Alake", "Cesare Bernardis", "Paul Cayet", "Luca Engel", "Damien Hilloulin", "Sungpack Hong", "Allen Hosler", "Nickolas Kavantzas", "Ingo Kossyk", "Son Le", "Rhicheek Patra", "Kartik Talamadupula", "Valentin Venzin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c0c60fd93a470f7d42588777e91374c56867dffd81f0d3a45d607c35a586550", "submitted_at": "2026-07-14T18:06:53Z", "updated_at": "2026-07-14T18:06:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.13157v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "id": "arxiv:2607.13157", "identifiers": {"arxiv": "2607.13157"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.13157"}], "name": "Oracle Agent Memory as an Enterprise Memory Substrate for Long-Horizon AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-14", "url": "https://arxiv.org/abs/2607.13157"}
{"aliases": [], "authors": ["Tianyu Chen", "Chujia Hu", "Wenjie Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "20355b736cabece09476f4b24b8d8dbf622470c7d23d163d2b8d1dcc2e7a24d6", "submitted_at": "2026-07-15T08:38:16Z", "updated_at": "2026-10-08T06:32:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.13594v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.13594", "identifiers": {"arxiv": "2607.13594"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.13594"}], "name": "SAFETY SENTRY: Context-Aware Human Intervention via EXECUTE-ASK-REFUSE Routing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2607.13594"}
{"aliases": [], "authors": ["Wenjun Wang", "Yuchen Fang", "Fengrui Liu", "Zibo Liang", "Kai Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "35460902368253008c14a4049e2fe1dbcd3a03e724022f2a637c3f6f314454cd", "submitted_at": "2026-07-15T14:33:15Z", "updated_at": "2026-08-10T09:35:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.13884v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.13884", "identifiers": {"arxiv": "2607.13884"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.13884"}], "name": "Experience Memory Graph: One-Shot Error Correction for Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-10", "url": "https://arxiv.org/abs/2607.13884"}
{"aliases": [], "authors": ["Leitian Tao", "Baolin Peng", "Wenlin Yao", "Tao Ge", "Hao Cheng", "Mike Hang Wang", "Jianfeng Gao", "Sharon Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c8a3be8c428d55aa988f6ceb0a168602c8357c41ea81367407d321a7e28ba8c", "submitted_at": "2026-07-15T16:16:42Z", "updated_at": "2026-07-15T16:16:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.13988v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.13988", "identifiers": {"arxiv": "2607.13988"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.13988"}], "name": "TRACE: Turn-level Reward Assignment via Credit Estimation for Long-Horizon Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-15", "url": "https://arxiv.org/abs/2607.13988"}
{"aliases": [], "authors": ["Amirhosein Ghasemabadi", "Ruichen Chen", "Bahador Rashidi", "Di Niu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b0d7133dd4505dbbcc594827be358ecf5e5225f617297b08767742088c70419", "submitted_at": "2026-07-15T18:36:54Z", "updated_at": "2026-07-15T18:36:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.14277v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.14277", "identifiers": {"arxiv": "2607.14277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.14277"}], "name": "Multi-Head Latent Control: A Unified Interface for LLM Agent Decision Making", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-15", "url": "https://arxiv.org/abs/2607.14277"}
{"aliases": [], "authors": ["Yong Liu", "Zhenyi Zhong", "Zhanpeng Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a779b2c1ff9ba3b531d399204b0986e3affcea81c67fdd7951e3b7da343b0e17", "submitted_at": "2026-07-16T00:35:25Z", "updated_at": "2026-07-16T00:35:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.14443v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.14443", "identifiers": {"arxiv": "2607.14443"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.14443"}], "name": "Tactile: Giving Computer-Using Agents Hands and Feet", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-07-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Tactile turns accessibility, OCR and visual evidence into explicit desktop targets with executable actions and verification cues. Its execution loop chooses semantic actions or coordinate fallbacks while retaining provenance for replay and diagnosis.", "topics": ["grounding", "automation", "observability", "tool-integration", "robustness", "evaluation"], "updated_at": "2026-07-16", "url": "https://arxiv.org/abs/2607.14443"}
{"aliases": [], "authors": ["Hy Vision Team", "Huawen Shen", "Zhengyang Tang", "Shangpin Peng", "Liang Wu", "Anran Zhang", "Weinong Wang", "Yiduo Guo", "Chenxin Li", "Zhengyao Fang", "Yang Ding", "Junyi Li", "Fei Tang", "Zheng Ruan", "Yi Zhang", "Xingran Zhou", "Dingchen Yang", "Sunqi Fan", "Zhiyi Wan", "Han Hu", "Xin Lai", "Pengyuan Lyu", "Chengquan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "151d4882fecc835dfffb9651bd860ebe8c85919a63ff5730e191f9146fbba272", "submitted_at": "2026-07-16T04:12:42Z", "updated_at": "2026-07-16T04:12:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.14548v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.14548", "identifiers": {"arxiv": "2607.14548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.14548"}], "name": "HyMobileAgent: Data-Environment Co-Scaling for Efficient GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines GUI perception data, instructional videos, real and simulated devices, and planning with reflection to develop a mobile agent.", "topics": ["mobile-navigation", "data-generation", "planning"], "updated_at": "2026-07-16", "url": "https://arxiv.org/abs/2607.14548"}
{"aliases": [], "authors": ["Madhumitha Venkatesan", "Shicheng Wen", "Jiajing Guo", "Jorge Piazentin Ono", "Liu Ren", "Dongyu Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1179c657434761187f3fa3bed99f46530a69878578b7a74c68febae20e64d6b3", "submitted_at": "2026-07-16T16:48:39Z", "updated_at": "2026-07-16T16:48:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.15193v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.15193", "identifiers": {"arxiv": "2607.15193"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.15193"}], "name": "Plover: Steering GUI Agents through Plan-Centric Interaction", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Plover exposes GUI-agent plans and replanning as persistent artifacts that users can inspect and edit during execution. Its planner–executor design supports localized correction through plan edits, language guidance and screenshot-grounded interventions, studied through failure repair and workflow scenarios.", "topics": ["planning", "human-in-the-loop", "observability", "workflow"], "updated_at": "2026-07-16", "url": "https://arxiv.org/abs/2607.15193"}
{"aliases": [], "authors": ["Xue Yu", "Bo Yuan", "Kailin Zhao", "Pengshuai Yang", "Hong Hu", "Junlan Feng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a15963a7fb4dfd488f0e79732d9ea0c9a03d43c1e3d5e50dd1ee8e00f1072be", "submitted_at": "2026-07-17T01:45:54Z", "updated_at": "2026-09-02T01:49:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.15550v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.15550", "identifiers": {"arxiv": "2607.15550"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.15550"}], "name": "SeerGuard: A Safety Framework for Mobile GUI Agents via World Model Prediction", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses instruction screening and a safety-augmented world model to assess likely consequences of mobile GUI actions before execution.", "topics": ["safety-security", "world-models"], "updated_at": "2026-09-02", "url": "https://arxiv.org/abs/2607.15550"}
{"aliases": [], "authors": ["Vishal Pandey", "Gopal Singh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e53594300b6be4e76f02f2c78906faa5999e727439067e6001c398eb7d97184", "submitted_at": "2026-07-17T12:12:38Z", "updated_at": "2026-07-17T12:12:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.15899v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.15899", "identifiers": {"arxiv": "2607.15899"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.15899"}], "name": "ContinuityBench: A Benchmark and Systems Study of Stateful Failover in Multi-Provider LLM Routing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-17", "url": "https://arxiv.org/abs/2607.15899"}
{"aliases": [], "authors": ["Xiaoye Zhu", "Weixin Li", "Junan Huo", "Bozhong Wang", "Jia Zeng", "Yi Yang", "Cen Chen", "Qi Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5b704f7f2d7f1a8966f2ed45045d9760e160c1f332b36fcb08909b8479e83482", "submitted_at": "2026-07-17T07:10:46Z", "updated_at": "2026-07-17T07:10:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.16352v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.16352", "identifiers": {"arxiv": "2607.16352"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.16352"}], "name": "Clarify Before Executing: A Self-Evolving Agent for Resolving Intent Asymmetry in 3D Tool Orchestration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-17", "url": "https://arxiv.org/abs/2607.16352"}
{"aliases": [], "authors": ["Yang Liu", "Weixing Chen", "Xinshuai Song", "Tao Pu", "Siwen Mo", "Yongjie Bai", "Zihao Chen", "Qianran Sun", "Liruo Zhong", "Ying Shen", "Liang Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "374857ccba8c7594441b358fc50775c2d1d90dec7b49bf69aec9a157a834d9f5", "submitted_at": "2026-07-18T04:46:53Z", "updated_at": "2026-07-18T04:46:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.16636v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.16636", "identifiers": {"arxiv": "2607.16636"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.16636"}], "name": "PhyAgentOS: A Self-Evolving Operating System for Embodied Agents with Decoupled Cognitive Planning and Physical Execution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-18", "url": "https://arxiv.org/abs/2607.16636"}
{"aliases": [], "authors": ["Seanie Lee", "Sanjoy Chowdhury", "Chao Jiang", "Cheng-Yu Hsieh", "Ting-Yao Hu", "Alexander T Toshev", "Oncel Tuzel", "Raviteja Vemulapalli"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ed498db51c7a93dc49d4d31754ab4a7814b2da62f45d423512b0338cf485da6c", "submitted_at": "2026-07-18T17:36:44Z", "updated_at": "2026-09-22T05:07:30Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.16900v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.16900", "identifiers": {"arxiv": "2607.16900"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.16900"}], "name": "Simulate to Generalize: Scaling Stateful Supervision for API-calling Agents using LLM World Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-22", "url": "https://arxiv.org/abs/2607.16900"}
{"aliases": [], "authors": ["Arunabh Dastidar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0fbdfba4b3cc1b6dd1c123ed5966118e8ef2ff98d10d669a0489317789cef7c1", "submitted_at": "2026-07-19T03:20:33Z", "updated_at": "2026-07-19T03:20:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.17044v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.17044", "identifiers": {"arxiv": "2607.17044"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.17044"}], "name": "Where Does Agent Reliability Come From? A Cross-Benchmark Decomposition of Verification Loops, Specialist Models, and Scaffolding in a Production Enterprise Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-19", "url": "https://arxiv.org/abs/2607.17044"}
{"aliases": [], "authors": ["Yaohan Yang", "Minglei Shi", "Borui Zhang", "Jie Zhou", "Jiwen Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6488798626abbec89ca739a0c5d188ae440484eb59b3be11f2b319ce742b22ec", "submitted_at": "2026-07-19T03:29:08Z", "updated_at": "2026-07-19T03:29:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.17050v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.17050", "identifiers": {"arxiv": "2607.17050"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.17050"}], "name": "EvoGUI: An Evolution-Aware Benchmark for GUI State-Transition Understanding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces EvoGUI, which derives temporal ordering, inverse action/value prediction, and successor-discrimination probes from recorded GUI trajectories. Its web-derived benchmark isolates transition understanding from the other abilities involved in end-to-end task success.", "topics": ["evaluation", "world-models", "action-prediction"], "updated_at": "2026-07-19", "url": "https://arxiv.org/abs/2607.17050"}
{"aliases": [], "authors": ["Ruei-Che Chang", "Wenqian Xu", "Dingzeyu Li", "Bryan Wang", "Anhong Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2069acf693e0dced49d0e10e21bf9fe28b99f50f1b932837f16137e0d3170a80", "submitted_at": "2026-07-20T04:03:38Z", "updated_at": "2026-07-20T04:03:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.17527v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.17527", "identifiers": {"arxiv": "2607.17527"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.17527"}], "name": "Sidekick: Designing Communication for Effective Multitasking with Computer Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Sidekick, a communication interface that uses ambient status cues, multimodal action summaries, and visual and verbal explanations of reasoning to help people monitor computer-use agents. A user study examines progress awareness, error traceability, and multitasking across background and foreground interaction.", "topics": ["human-in-the-loop", "observability", "long-horizon"], "updated_at": "2026-07-20", "url": "https://arxiv.org/abs/2607.17527"}
{"aliases": [], "authors": ["Di Lu", "Bo Zhang", "Xiyuan Li", "Yongzhi Liao", "Xuewen Dong", "Yulong Shen", "Zhiquan Liu", "Jianfeng Ma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a8253e94ce43c0a614d4f56ad4e68c574b174de62d3ac6dafcd8bc2f360aced", "submitted_at": "2026-07-20T13:52:24Z", "updated_at": "2026-07-20T13:52:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.17951v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.17951", "identifiers": {"arxiv": "2607.17951"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.17951"}], "name": "RT-SHCUA: Real-Time Self-Hosted Computer-Use Agent for UAV Control", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-20", "url": "https://arxiv.org/abs/2607.17951"}
{"aliases": [], "authors": ["Zhuohang Fan", "Beichen Zhang", "Yuanfa Li", "Changqiao Wu", "Wei Liu", "Jian Luan", "Weigang Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b85b453a688543477d31a940f9d1da2857b2cb7a8cb9eea91a39a3a2b8a1e1c7", "submitted_at": "2026-07-20T15:16:59Z", "updated_at": "2026-08-18T02:32:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.18046v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.18046", "identifiers": {"arxiv": "2607.18046"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.18046"}], "name": "SEE: Structure-aware Exploring & Exploiting for Long-horizon GUI Agent Trajectory Synthesis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SEE, which first explores mobile interfaces to build a screen-and-element transition graph, then composes task trajectories through graph planning and controlled sampling. It studies coverage, loop avoidance, and transfer after training agents on the generated sequences.", "topics": ["data-generation", "long-horizon", "mobile-navigation"], "updated_at": "2026-08-18", "url": "https://arxiv.org/abs/2607.18046"}
{"aliases": [], "authors": ["Behzad Ousat", "Nikita Turkmen", "Lalchandra Rampersaud", "Dillan Bailey", "Amin Kharraz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e382b15a9d25e99aa187a72c7d2a33bbeca82b2b347ba580ca0a1d73a8e01d19", "submitted_at": "2026-07-21T03:02:57Z", "updated_at": "2026-07-21T03:02:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.18659v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2607.18659", "identifiers": {"arxiv": "2607.18659"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.18659"}], "name": "Broken Gates: Re-evaluating Web Bot Defenses in the Age of LLM Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Measures how autonomous LLM browser agents interact with commercial bot-management and anti-automation defenses on the web.", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2026-07-21", "url": "https://arxiv.org/abs/2607.18659"}
{"aliases": [], "authors": ["Jiazhen Jiang", "Boxi Cao", "Lingyong Yan", "Yaojie Lu", "Hongyu Lin", "Shuaiqiang Wang", "Dawei Yin", "Xianpei Han", "Le Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "83e4346ec3a6b7e2de767fd4bb18df73288172d8eac4b9504684223dd45b2e3b", "submitted_at": "2026-07-22T07:52:59Z", "updated_at": "2026-07-22T07:52:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.19865v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.19865", "identifiers": {"arxiv": "2607.19865"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.19865"}], "name": "DocOps: A Verifiable Benchmark for Autonomous Agents in Complex Document Operations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-22", "url": "https://arxiv.org/abs/2607.19865"}
{"aliases": [], "authors": ["Alexis Fox", "Junlin Wang", "Paul Rosu", "Bhuwan Dhingra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "664004d0e93848071ef7028c785b243406ce0e5f5b577e212df0ba34b1e9c143", "submitted_at": "2026-07-22T12:11:51Z", "updated_at": "2026-10-12T06:47:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.20064v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.20064", "identifiers": {"arxiv": "2607.20064"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.20064"}], "name": "PRO-LONG: Programmatic Memory Enables Long-Horizon Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2607.20064"}
{"aliases": [], "authors": ["Qiyuan Liu", "Tingfeng Hui", "Kun Zhan", "Kaike Zhang", "Ning Miao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f74668d24cf77d29be8742b6b75c8825a5a43dfefbac167df9dca663c28b5a7d", "submitted_at": "2026-07-22T13:24:09Z", "updated_at": "2026-07-23T09:17:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.20121v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.20121", "identifiers": {"arxiv": "2607.20121"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.20121"}], "name": "OpenSkillRisk: Benchmarking Agent Safety When Using Real-World Risky Third-Party Skills", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-23", "url": "https://arxiv.org/abs/2607.20121"}
{"aliases": [], "authors": ["Seungbin Yang", "Chaewoon Ki", "Dohyun Lee", "Jaegul Choo", "ChaeHun Park"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "36e975245dab12dcbf586a7bc4ff5fca95402d58740146d210da37d4e6d6afe0", "submitted_at": "2026-05-30T13:27:55Z", "updated_at": "2026-08-04T07:30:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.20482v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.20482", "identifiers": {"arxiv": "2607.20482"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.20482"}], "name": "PersonaTrail: Benchmarking Personalized Web Agents through Browsing Trails", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-05-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2607.20482"}
{"aliases": [], "authors": ["Hongxin Zhang", "Chunru Lin", "Junyan Li", "Zhou Xian", "Tsun-Hsuan Wang", "Chuang Gan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "634ca02f1823aebaf1d56952caf3fddb6d197a947d6ffc84adc2a96a9c5df5d2", "submitted_at": "2026-07-23T17:04:36Z", "updated_at": "2026-07-23T17:04:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.21522v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.21522", "identifiers": {"arxiv": "2607.21522"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.21522"}], "name": "GS-Agent: Creating 4D Physical Worlds With Generative Simulation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-23", "url": "https://arxiv.org/abs/2607.21522"}
{"aliases": [], "authors": ["Xiao Yu", "Baolin Peng", "Ruize Xu", "Hao Zou", "Qianhui Wu", "Hao Cheng", "Wenlin Yao", "Nikhil Singh", "Zhou Yu", "Jianfeng Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "965fe5aabfb44b4c82290fc347a463eb58fb2be7a3761496efd38167d0f7d5b6", "submitted_at": "2026-07-23T17:38:30Z", "updated_at": "2026-08-07T17:40:50Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.21557v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.21557", "identifiers": {"arxiv": "2607.21557"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.21557"}], "name": "OpenForgeRL: Train Harness-native Agents in Any Environment", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "OpenForgeRL connects stateful agent harnesses to reinforcement-learning infrastructure through a model-call recording proxy and isolated rollout containers. Its validation includes browser and computer-use agents, alongside tool-oriented agents, and examines how harness choice affects learned reliability.", "topics": ["reinforcement-learning", "sandbox", "tool-integration", "robustness"], "updated_at": "2026-08-07", "url": "https://arxiv.org/abs/2607.21557"}
{"aliases": [], "authors": ["Zedong Yu", "Qianxing Li", "Zhi Gao", "Liuyu Xiang", "Chenrui Shi", "Yang Liu", "Huiming Wu", "Yujie Wei", "Yuhao Fei", "Yubo Fu", "Zhaofeng He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "078ff9a56b769923d15a98581685a3e2f47aac7d898ea731999cabaec3918fa7", "submitted_at": "2026-07-17T12:48:03Z", "updated_at": "2026-07-17T12:48:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.22689v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.22689", "identifiers": {"arxiv": "2607.22689"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.22689"}], "name": "Beyond Sequential Interaction: Benchmarking Parallel Execution and Coordination for GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ParaGUIBench, with tasks and efficiency measurements for coordinated GUI agents operating separate desktop instances over a shared file system. Its ParaGUI planner-worker system decomposes tasks and dispatches concurrent subtasks for comparison with serial execution.", "topics": ["evaluation", "planning", "efficiency", "long-horizon"], "updated_at": "2026-07-17", "url": "https://arxiv.org/abs/2607.22689"}
{"aliases": [], "authors": ["Nimra Noor", "Muhammad Bilal", "Abdullah Hussain", "Hassan Baig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e558081cd9d7c66ade64bf07054c18a995fc4794de3668e1aba799f9916148b3", "submitted_at": "2026-07-22T16:33:54Z", "updated_at": "2026-07-22T16:33:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.22738v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.22738", "identifiers": {"arxiv": "2607.22738"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.22738"}], "name": "Nova3D: Code-Native Generation of Programmable 3D Assets", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-22", "url": "https://arxiv.org/abs/2607.22738"}
{"aliases": [], "authors": ["Yan Yang", "Xiangru Jian", "Ziyang Luo", "Zirui Zhao", "Yutong Dai", "Ziji Shi", "Hanshu Yan", "Jun Hao Liew", "Silvio Savarese", "Junnan Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a010fbef09d8359550cf156bf0dcba873f25a0d6fb8f118d53bf6e23715cf203", "submitted_at": "2026-07-24T14:17:03Z", "updated_at": "2026-07-24T14:17:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.22798v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.22798", "identifiers": {"arxiv": "2607.22798"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.22798"}], "name": "StateAct: Program State, before Pixels, for Long-Horizon Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces StateAct, a code-first computer-use harness that manipulates program state directly and delegates screenshot-and-click subgoals to a GUI agent. An independent completion gate checks saved outputs, while fresh subagents handle subgoals to limit accumulated context.", "topics": ["tool-integration", "state-tracking", "long-horizon"], "updated_at": "2026-07-24", "url": "https://arxiv.org/abs/2607.22798"}
{"aliases": [], "authors": ["Yang Wan", "Zhenhao Zhang", "Jierui Wang", "Xiangxin Zhou", "Linchao Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2de35aa3c7f4457d43d485eb136d8b380ea61870cc7253e807880ae731c51316", "submitted_at": "2026-07-25T16:00:45Z", "updated_at": "2026-09-26T12:15:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.23263v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.23263", "identifiers": {"arxiv": "2607.23263"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.23263"}], "name": "SeekJudge: A Practical Reward Framework for Reinforcement Learning in Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SeekJudge, which iteratively searches and analyzes GUI trajectories to assess instruction fulfillment. The work combines dense judgment-data synthesis, a step-label benchmark, and reward-server scheduling to study accurate and practical reward feedback for agent training.", "topics": ["reward-models", "evaluation", "data-generation"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2607.23263"}
{"aliases": [], "authors": ["Weihuang Zheng", "Tianyuan Zou", "Eileen Ye", "Alphet Liu", "Youyong Kong", "Ya-Qin Zhang", "Duran Zheng", "Maxm Pan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "235193576f6e3a89a393af538e1a84db1ec0459a919968c83c5bdbe019a45e43", "submitted_at": "2026-07-26T15:38:28Z", "updated_at": "2026-07-26T15:38:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.23722v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.23722", "identifiers": {"arxiv": "2607.23722"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.23722"}], "name": "E-Bench: Benchmarking Multi-Step Tool-Use Agents in Real-World Product Scenarios", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-26", "url": "https://arxiv.org/abs/2607.23722"}
{"aliases": [], "authors": ["Xiangyan Liu", "Kaixin Li", "Haonan Wang", "Biao Wu", "Meng Fang", "Longxu Dou", "Chao Du", "Michael Qizhe Shieh", "Tianyu Pang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4651964cbe45e4d740ed8f789753f82ae808e442d82c6598e7b3c997c09453c3", "submitted_at": "2026-07-27T07:54:08Z", "updated_at": "2026-07-27T07:54:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.24112v1"}}, "environments": ["desktop", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.24112", "identifiers": {"arxiv": "2607.24112"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.24112"}], "name": "Scaling GUI Agents with Visual State Transitions", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces State Transition Pretraining, which jointly learns to infer actions from GUI changes and predict future states from current screens and actions. Subsequent instruction-trajectory fine-tuning tests whether these representations transfer to desktop and mobile agent tasks.", "topics": ["world-models", "action-prediction", "data-generation"], "updated_at": "2026-07-27", "url": "https://arxiv.org/abs/2607.24112"}
{"aliases": [], "authors": ["Guangyi Liu", "Huan Zhao", "Quanming Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57fa0a3eaad7525991656989ed3011101026362acec43895f32cf6e24be2f4ee", "submitted_at": "2026-07-27T08:48:01Z", "updated_at": "2026-07-27T08:48:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.24167v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.24167", "identifiers": {"arxiv": "2607.24167"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.24167"}], "name": "Falsifiable Commitment Planning for Self-Correcting Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces FCPAgent, which attaches confirming and falsifying evidence to each skill-grounded plan step. A plan–test–repair loop checks actions and observations, then localizes contradictions so execution, skills, or plans can be revised at the appropriate level.", "topics": ["planning", "robustness", "self-reflection"], "updated_at": "2026-07-27", "url": "https://arxiv.org/abs/2607.24167"}
{"aliases": [], "authors": ["Tianyi Men", "Zhuoran Jin", "Kang Liu", "Jun Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c94a1a2780ae2ffe6acce3b8388ed55c45e454200a91fab6975da076716969a9", "submitted_at": "2026-07-27T17:55:03Z", "updated_at": "2026-07-27T17:55:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.24720v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.24720", "identifiers": {"arxiv": "2607.24720"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.24720"}], "name": "The Physics of Multi-Turn Long-Horizon Planning: From Pre-training to Post-training via Single- and Multi-Teacher On-Policy Agentic Distillation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-27", "url": "https://arxiv.org/abs/2607.24720"}
{"aliases": [], "authors": ["Haowen Dai", "Zonghao Ying", "Wenfeng Li", "Xiangfan Wu", "Yisong Xiao", "Tianyuan Zhang", "Jiaye Lin", "Lei Wei", "Guangyuan Dong", "Xitong Ling", "Xixun Lin", "Quanchen Zou", "Xiangzheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c6a6d103ae967adb0935497b7b02cf11f7ab9b45eff3513c9ec30cff6ae5e04f", "submitted_at": "2026-07-28T03:55:47Z", "updated_at": "2026-07-29T23:58:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.25255v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2607.25255", "identifiers": {"arxiv": "2607.25255"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.25255"}], "name": "SafeFlow: Semantic Information-Flow Control for Blocking Malicious Propagation in Multi-Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-29", "url": "https://arxiv.org/abs/2607.25255"}
{"aliases": [], "authors": ["Chenrui Shi", "Yuwei Wu", "Yang Liu", "Ruining Feng", "Zirui Shang", "Zhi Gao", "Lifeng Fan", "Che Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d1688a528d35a8213a8027385b1b1327536567f6ec307aa6562eea543ee71d7f", "submitted_at": "2026-07-28T16:01:38Z", "updated_at": "2026-07-29T05:13:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.25904v2"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.25904", "identifiers": {"arxiv": "2607.25904"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.25904"}], "name": "Interactive Reward Agent: GUI Task Evaluation via Environment-State Verification", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces an Interactive Reward Agent that proposes completion conditions and verifies them using system, application, and GUI tools after execution. GUI-RewardBench provides annotated desktop trajectories for evaluating these evidence-based judgments and their use in reinforcement learning.", "topics": ["reward-models", "evaluation", "tool-integration"], "updated_at": "2026-07-29", "url": "https://arxiv.org/abs/2607.25904"}
{"aliases": [], "authors": ["Abhishek Pillai", "Samir Kumar Nayak", "Yuan Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "352c892a52692301e14125e126a52508feb39ef7cd034d7cb0c48a27fb54b8ac", "submitted_at": "2026-07-28T17:49:51Z", "updated_at": "2026-07-29T19:07:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.26041v2"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.26041", "identifiers": {"arxiv": "2607.26041"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.26041"}], "name": "Desktop-Delta Bench: Do Computer-Use Models Understand Desktop GUI Transitions?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Desktop-Delta Bench, an offline diagnostic of GUI transition understanding using temporal-ordering tasks and before/after action inference. It includes decoy observations to test whether models track the source and task relevance of screen changes.", "topics": ["evaluation", "screen-understanding", "state-tracking"], "updated_at": "2026-07-29", "url": "https://arxiv.org/abs/2607.26041"}
{"aliases": [], "authors": ["Xuan Zhao", "Jiwoong Sohn", "Qinyue Zheng", "Michael Moor"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4acd560354e9a7967e29761f3a008c822d965b32ae3072ba22482958f2ff274e", "submitted_at": "2026-07-28T21:47:22Z", "updated_at": "2026-08-04T13:59:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.26300v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.26300", "identifiers": {"arxiv": "2607.26300"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.26300"}], "name": "AgentGUI: An Interface for Observing and Steering Long-Running AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2607.26300"}
{"aliases": [], "authors": ["Siyu Yan", "Zhuoran Yan", "Haiying Xu", "Panhao Zhou", "Jingyu Chen", "Chenhao Ji", "Shuo Cao", "Yongheng Zhang", "Haoze Liu", "Siyu Zhang", "Xiwen Gu", "Yihao Liu", "Alex Jinpeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb11fc857750d5a67e9bddc8a28e20ba7301ae8c660c08f9af2bd7a573d6a0f1", "submitted_at": "2026-07-29T11:10:34Z", "updated_at": "2026-07-29T11:10:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.26769v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.26769", "identifiers": {"arxiv": "2607.26769"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.26769"}], "name": "See2Think: Do Multimodal Models Really Use Intermediate Visual States?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-29", "url": "https://arxiv.org/abs/2607.26769"}
{"aliases": [], "authors": ["Jingbo Zhou", "Yusai Zhao", "Qi Bao", "Jingjia Cao", "Zhenghai Chen", "Chang Gao", "Kaiqi Guo", "Muxin Guo", "Mingxuan Li", "Xinjiang Lu", "Yanru Ma", "Yixiong Xiao", "Zenghui Zhang", "Le Zhang", "Hua Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e743f464b2ca509f50d372e0f9dc9a49397ffb6f4d24be66452284557c97ad6", "submitted_at": "2026-07-29T17:33:47Z", "updated_at": "2026-08-18T12:27:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.27155v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2607.27155", "identifiers": {"arxiv": "2607.27155"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.27155"}], "name": "OmegaUse-OfficeVal: Benchmarking LLM Agents on Long-Horizon Office-Suite Tasks with Economic Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-18", "url": "https://arxiv.org/abs/2607.27155"}
{"aliases": [], "authors": ["Nianchen Deng", "Jiaxin Ai", "Tao Hu", "Shu Zou", "Yurui Dong", "Siqi Li", "Xinyu Cai", "Xuemeng Yang", "Licheng Wen", "Hongbin Zhou", "Hairong Zhang", "Pinlong Cai", "Botian Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "99c9308a622318b8ddd77cde6e843746e4039483f3e1e89f27bc99738890f003", "submitted_at": "2026-07-30T11:28:21Z", "updated_at": "2026-07-30T11:28:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28050v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-19T15:37:15Z", "role": "primary", "sha256": "4d02c402f4cadeee710c838f7edfd2c720346dc1afc54d9a3f1b75736035954b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.28050&max_results=1"}], "id": "arxiv:2607.28050", "identifiers": {"arxiv": "2607.28050"}, "kind": "paper", "last_verified_at": "2026-10-19T15:37:15Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.28050"}], "name": "IndustryForge-27B: A Domain-Enhanced Multimodal Foundation Model for Industrial CAD", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-30", "url": "https://arxiv.org/abs/2607.28050"}
{"aliases": [], "authors": ["Yash Pandya", "Sahil Gupta", "Sarthak Harne", "Archana Yadav", "Kavyansh Chourasia", "Hussein Mozannar", "Vibhav Vineet", "Sara Abdali", "Corby Rosset", "Yash Lara", "Ahmed Awadallah", "Ece Kamar", "Akshay Nambi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f28e1ee8e101000f3cfe9215555251656fcb40b921b5c41c895fbc958f867e3", "submitted_at": "2026-07-30T11:48:22Z", "updated_at": "2026-07-30T11:48:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28074v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.28074", "identifiers": {"arxiv": "2607.28074"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28074"}], "name": "Echoverse: Deep, Evolving Environments for Training Computer-Use Agents at Scale", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Echoverse, which compiles specifications into stateful applications with database-grounded graders and uses rollouts to improve both the environments and agent policies. It examines behavioral depth, targeted interface practice, and environment repair as sources of training value.", "topics": ["sandbox", "data-generation", "reinforcement-learning"], "updated_at": "2026-07-30", "url": "https://arxiv.org/abs/2607.28074"}
{"aliases": [], "authors": ["Hanzhang Zhou", "Panrong Tong", "Xu Zhang", "Quyu Kong", "Chenglin Cai", "Tianyu Xia", "Gongjie Zhang", "Jianan Zhang", "Long Li", "Long Chen", "Lei Wang", "Gaole Dai", "Pengxiang Li", "Liangyu Chen", "Yue Wang", "Steven Hoi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52fb1b49501132f1dbe3564daa02cca22a93fc8d7c28dec8ee6070a23889574a", "submitted_at": "2026-07-30T13:58:41Z", "updated_at": "2026-07-30T13:58:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28227v1"}}, "environments": ["cross-platform", "web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.28227", "identifiers": {"arxiv": "2607.28227"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28227"}], "name": "Qwen-UI-Agent Technical Report: Toward Next-Generation Real-World Centric Foundation GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a foundation GUI agent with combined GUI and command-line actions, large-scale interaction infrastructure, and iterative environment and task generation.", "topics": ["reinforcement-learning", "long-horizon", "tool-integration"], "updated_at": "2026-07-30", "url": "https://arxiv.org/abs/2607.28227"}
{"aliases": [], "authors": ["Luigi Sigillo", "Matteo Silvestri", "Francesco Tabaro", "Rajat Bhatnagar", "Syed Irtaza Mubashar", "Matt Jeffryes", "Daljit Nijjer", "Vittorio Perera", "Ola Spjuth", "Julio Saez-Rodriguez", "Melissa Harrison", "Fabio Petroni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b0585630011697d8aa009a306d01284be9e1f7bc7e4aeac8331cda6c82915399", "submitted_at": "2026-07-30T14:00:27Z", "updated_at": "2026-07-30T14:00:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28229v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.28229", "identifiers": {"arxiv": "2607.28229"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.28229"}], "name": "EMBL AI Librarian: Life-Sciences Knowledge Layer for AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-30", "url": "https://arxiv.org/abs/2607.28229"}
{"aliases": [], "authors": ["Rong Wu", "Daocheng Fu", "Licheng Wen", "Xuemeng Yang", "Shu Zou", "Jianbiao Mei", "Yuxin Wang", "Hairong Zhang", "Yu Yang", "Tao Hu", "Cong Zhang", "Botian Shi", "Pinlong Cai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6470315f4b51c0dcd2cde5e1ea46c5b4cc873f5d4ab7ea462c518c7aec5d0fd2", "submitted_at": "2026-07-30T14:25:49Z", "updated_at": "2026-07-30T14:25:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28272v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.28272", "identifiers": {"arxiv": "2607.28272"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.28272"}], "name": "MemHarness: Memory Is Reconstructed, Not Replayed", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-30", "url": "https://arxiv.org/abs/2607.28272"}
{"aliases": [], "authors": ["Zihan Dong", "Zhiyuan Ma", "Zekun Wang", "Yunqing Li", "Zirou Liu", "Ruixuan Deng", "Qishi Zhan", "Rui Qian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d75f6c942d6854a259556b7ef1060fa0f142755162da329580c6177cd1adef15", "submitted_at": "2026-07-30T15:29:42Z", "updated_at": "2026-07-30T15:29:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28367v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.28367", "identifiers": {"arxiv": "2607.28367"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28367"}], "name": "How Benchmarks Mis-Score Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Audits failure-scored computer-use trajectories to distinguish agent errors from broken tasks and evaluator false negatives. It organizes reliability issues across task construction, trajectory evidence, scoring, and reporting, then derives corresponding evaluation design practices.", "topics": ["evaluation", "observability", "robustness"], "updated_at": "2026-07-30", "url": "https://arxiv.org/abs/2607.28367"}
{"aliases": [], "authors": ["Zihan Dong", "Rui Qian", "Qishi Zhan", "Dongshen Peng", "Kaixin Li", "Yu Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "463cf35eb57b3d38dda6ce768e6f08d93b1c16d6abba0d790760d846cd56fa2c", "submitted_at": "2026-07-30T15:50:10Z", "updated_at": "2026-07-30T15:50:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28399v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.28399", "identifiers": {"arxiv": "2607.28399"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28399"}], "name": "Why Are GUI Agents Correct but Late? Decode on the Decision-Time Critical Path, Tested with Pre-Compiled Policy Trees", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Adaptive Anticipatory Policy Trees, which precompute guarded actions and deadlines during idle periods so transient GUI events can be handled without new text generation. Paired experiments examine when this prepared execution helps and when reactive control remains preferable.", "topics": ["efficiency", "planning", "robustness"], "updated_at": "2026-07-30", "url": "https://arxiv.org/abs/2607.28399"}
{"aliases": [], "authors": ["Xinxing Ren", "Qianbo Zang", "Ziyan Wang", "Caelum Forder", "Suman Deb", "Peter Carroll", "Zekun Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "48d0a6b4dfb301f6f38ad87ccf874d7e483b4b5fd232b55e02699f387b78abdf", "submitted_at": "2026-07-30T16:07:32Z", "updated_at": "2026-07-30T16:07:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28430v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2607.28430", "identifiers": {"arxiv": "2607.28430"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.28430"}], "name": "AgentRadio: Passive Awareness for Long-Horizon Multi-Agent Collaboration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-30", "url": "https://arxiv.org/abs/2607.28430"}
{"aliases": [], "authors": ["Woongkyu Lee", "Jungwook Choi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bacd1c5df5bf3163c83fae546ae4ced36996f33115568dfbc9e44b17fe507a31", "submitted_at": "2026-07-30T17:36:36Z", "updated_at": "2026-07-30T17:36:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28573v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.28573", "identifiers": {"arxiv": "2607.28573"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28573"}], "name": "Rethinking Inference-Time Scaling in Local Computer-Use Agents: Failure Modes and Compute Tradeoffs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Examines inference-time scaling for local computer-use models by varying context, step budgets, decomposition, and parallel execution. OSWorld experiments track both task performance and changing failure modes, exposing cases where extra computation extends errors or adds coordination overhead.", "topics": ["efficiency", "evaluation", "robustness"], "updated_at": "2026-07-30", "url": "https://arxiv.org/abs/2607.28573"}
{"aliases": [], "authors": ["Qiushi Sun", "Kanzhi Cheng", "Yian Wang", "Bowen Yang", "Hang Yan", "Liheng Chen", "Fangzhi Xu", "Zichen Ding", "Nuo Chen", "Jialin Cao", "Xingdong Gong", "Zehao Li", "Kaiming Jin", "Xinfeng Yuan", "Zhoumianze Liu", "Jingyang Gong", "Zhangyue Yin", "Jiahui Gao", "Zhiyong Wu", "Tianbao Xie", "Jianbing Zhang", "Ben Kao", "Lingpeng Kong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3e82df7abffbd0c8f7e969ed8edb77423263a2b1d826a3e72529b0a51903d58", "submitted_at": "2026-07-30T17:57:41Z", "updated_at": "2026-08-06T17:55:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.28609v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.28609", "identifiers": {"arxiv": "2607.28609"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28609"}], "name": "OSReward: Instituting Standardized Evaluation for Cross-Platform Computer-Use Reward Models", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces OSReward and its harder and multidimensional variants for assessing judgments of computer-use trajectories. It also releases a reasoning-annotated judgment corpus and trains reward models, examining reliability, leniency bias, and evaluation cost.", "topics": ["reward-models", "evaluation", "data-generation"], "updated_at": "2026-08-06", "url": "https://arxiv.org/abs/2607.28609"}
{"aliases": [], "authors": ["Haoxin An", "Yunpeng Song", "Zihao Bai", "Zhongmin Cai", "Guojun Xiong", "Chenhao Lin", "Wentao Chen", "Feng Wei", "Chao Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45e0f558caad8e513c8672239dc04ce5f1557151aa984c6e9a138ad71c184a4c", "submitted_at": "2026-07-31T09:18:55Z", "updated_at": "2026-07-31T09:18:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.29199v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.29199", "identifiers": {"arxiv": "2607.29199"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2607.29199"}], "name": "Alignment Is Local: A Paired Diagnostic for GUI Agents under User-Side Persuasion", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-31", "url": "https://arxiv.org/abs/2607.29199"}
{"aliases": [], "authors": ["Hang Yan", "Zhangxuan GU", "Beitong Zhou", "Jiaxuan Chen", "Runze Li", "Yusong Hu", "Shuheng Shen", "Changhua Meng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfffaffc68695489e2a705515378112b232e66b6f9e8e3a9bd5df04be95bd2c9", "submitted_at": "2026-07-31T11:51:04Z", "updated_at": "2026-07-31T11:51:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.29320v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2607.29320", "identifiers": {"arxiv": "2607.29320"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.29320"}], "name": "MAGA: Multi-Platform Self-Fusion of GUI Agents via Structured Action Distillation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "MAGA distills specialized GUI agents into one policy by allocating supervision according to structured-action correctness rather than treating all response tokens equally. A training-only teacher hint supports fusion across mobile, web and desktop settings without changing student inputs.", "topics": ["imitation-learning", "grounding", "automation"], "updated_at": "2026-07-31", "url": "https://arxiv.org/abs/2607.29320"}
{"aliases": [], "authors": ["Yashaswi Malla", "Sandra Siby"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60c93c90e6d4026452810a866a0c49d57afdec15ad5ced30606c925a8e83c8f7", "submitted_at": "2026-07-31T18:27:12Z", "updated_at": "2026-07-31T18:27:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.00202v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2608.00202", "identifiers": {"arxiv": "2608.00202"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.00202"}], "name": "From Monoliths to Swarms: A Study of Attack Surface Evolution in the Transition to Multi-Agent Web Systems", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Categorizes structural attack surfaces that emerge when single-agent browser systems are decomposed into multi-agent web workflows.", "topics": ["safety-security", "web-navigation", "planning"], "updated_at": "2026-07-31", "url": "https://arxiv.org/abs/2608.00202"}
{"aliases": [], "authors": ["Chengbo Liu", "Lifang Zhou", "Ruijie Yan", "Pei Tan", "Ao Sun", "Haojun Huang", "Guichun Hua", "Sining Wei", "Yining Chen", "Yingying He", "Yutao Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1736defb4099677df47f882e02b6aad4066f8efa3065d76a1b2fb0f424d2fa19", "submitted_at": "2026-07-31T22:57:16Z", "updated_at": "2026-07-31T22:57:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.00335v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.00335", "identifiers": {"arxiv": "2608.00335"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.00335"}], "name": "RMSWeb: Reflection, Failure-Mode Mining, and Salvage-DS for Web Agent Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-31", "url": "https://arxiv.org/abs/2608.00335"}
{"aliases": [], "authors": ["Pu Cao", "Qingye Kong", "Xuedan Yin", "Xuekun Zhao", "Rupeng Yan", "Qing Song", "Yao Zhang", "Lu Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b55f44e80c7c4f8a733dc842133a48227e11379c1d2688d9b912429f29b22140", "submitted_at": "2026-08-01T09:05:06Z", "updated_at": "2026-08-01T09:05:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.00548v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.00548", "identifiers": {"arxiv": "2608.00548"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.00548"}], "name": "DrawAI: Agentic Benchmark and Workflow for Making Raster Images Editable", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-01", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-01", "url": "https://arxiv.org/abs/2608.00548"}
{"aliases": [], "authors": ["Ziyu Ma", "Hailang Huang", "Shun Zou", "Yong Wang", "Shidong Yang", "Yiming Hu", "Fei Wei", "XiangXiang Chu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9a8e00c323accb28a804e49ad903a47b92c01d12b9c7c3b8219ad752f8ba5883", "submitted_at": "2026-08-03T09:32:21Z", "updated_at": "2026-08-03T09:32:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.01964v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2608.01964", "identifiers": {"arxiv": "2608.01964"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.01964"}], "name": "LongHorizon-Harness: Advancing Long-Horizon Agents for Real-World Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-03", "url": "https://arxiv.org/abs/2608.01964"}
{"aliases": [], "authors": ["Jia-Chen Zhang", "Ze-Yu Zhang", "Kai-Wei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1111a54a66183f6f0ab0bbde067937c1806214821e5ed487a3eb3367c57050ce", "submitted_at": "2026-08-03T10:16:53Z", "updated_at": "2026-08-06T02:33:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.02018v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.02018", "identifiers": {"arxiv": "2608.02018"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.02018"}], "name": "Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-06", "url": "https://arxiv.org/abs/2608.02018"}
{"aliases": [], "authors": ["Can Wang", "Haoran Chen", "Haowen Gao", "Hao Ding", "Zhaoyang Liu", "Zhiying Tu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1346509ffe1bcb920a3e7216a85ca0a1a9ce45ce753e92b8346a8101334c029", "submitted_at": "2026-08-03T12:43:00Z", "updated_at": "2026-08-03T12:43:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.02163v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.02163", "identifiers": {"arxiv": "2608.02163"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.02163"}], "name": "From Simple QA to Deep Research: A Verifiable Benchmark Constructed through Iterative Task Evolution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-03", "url": "https://arxiv.org/abs/2608.02163"}
{"aliases": [], "authors": ["Dunjie Lu", "Shuai Bai", "Tianyi Bai", "Sicheng Fan", "Chang Gao", "Jian Guan", "Feng Hu", "Mianqiu Huang", "Xingyang Huang", "Yizhen Jiang", "Yuheng Jing", "Dehui Kong", "Ning Li", "Dayiheng Liu", "Shixuan Liu", "Zheng Liu", "Que Shen", "Bowen Wang", "Junli Wang", "Chencan Wu", "Rui Xie", "Tianbao Xie", "Zhihui Xie", "Haiyang Xu", "An Yang", "Tao Yu", "Wenzhen Yuan", "Xi Zhang", "Zhenru Zhang", "Mingkang Zhu", "Zhaoqing Zhu", "Yizhong Cao", "Kai Dang", "Binyuan Hui", "Kaixin Li", "Junyang Lin", "Haiquan Wang", "Zekun Wang", "Yiheng Xu", "Fan Yan", "Mengqi Yuan", "Danyang Zhang", "Jiajun Zhang", "Zhipeng Zhang", "Fan Zhou", "Fan Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "26a61ccfdddb5cdc2da2b14d29ae8d4098544fea4e9458126f2a49c28418c4f0", "submitted_at": "2026-08-03T15:04:20Z", "updated_at": "2026-08-03T15:04:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.02352v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.02352", "identifiers": {"arxiv": "2608.02352"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.02352"}], "name": "Qwen-CUA: Native Computer Use for (almost) Everything", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-08-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Qwen-CUA operates through screenshots and keyboard/mouse events while compressing older visual history into reusable context blocks. Its report combines verifiable task generation, trajectory-based training and separate browser and Bash-augmented studies.", "topics": ["reinforcement-learning", "imitation-learning", "data-generation", "interaction-history", "long-horizon", "evaluation"], "updated_at": "2026-08-03", "url": "https://arxiv.org/abs/2608.02352"}
{"aliases": [], "authors": ["Weijia Zhang", "Kunlun Zhu", "Zeyi Liu", "Yinting Chen", "Tianyi Ma", "Jiateng Liu", "Jiaxun Zhang", "Bingxuan Li", "Xiangru Tang", "Heng Ji", "Jiaxuan You"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a11e065bef4cf860c36ca2f682801ca21b552fbd40bd1b6f5d9ae39b56e7d77a", "submitted_at": "2026-07-31T13:59:45Z", "updated_at": "2026-09-06T04:43:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.02643v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.02643", "identifiers": {"arxiv": "2608.02643"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.02643"}], "name": "CUADebug: Diagnosing and Repairing Computer-Use Agent Failures", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-07-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "CUADebug combines a failure taxonomy and human-annotated OSWorld failure trajectories with an agent that locates causal steps and proposes corrections from screenshots and action traces. It evaluates root-cause diagnosis separately from controlled task re-execution and failure recovery.", "topics": ["evaluation", "robustness", "self-reflection", "state-tracking"], "updated_at": "2026-09-06", "url": "https://arxiv.org/abs/2608.02643"}
{"aliases": [], "authors": ["Ranxu Zhang", "Guinan Chen", "Chenshaodong", "Jinghao Lin", "Xiaozhou Xu", "Sunzhe", "Yanyong Zhang", "Chao Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea1d34cdfaa71558ac1e2bd7eef15de346ea3a7a8a5ebc5c1d80d9f5868e4878", "submitted_at": "2026-08-04T06:56:47Z", "updated_at": "2026-08-04T06:56:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.03223v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.03223", "identifiers": {"arxiv": "2608.03223"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.03223"}], "name": "Agentic Reinforcement Learning with Self-Distilled Reward Shaping", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2608.03223"}
{"aliases": [], "authors": ["Zichuan Fu", "Shirong Wang", "Wenlin Zhang", "Guojing Li", "Yimin Deng", "Jingtong Gao", "Junjia Qi", "Hanyu Yan", "Yefeng Zheng", "Xiaopeng Li", "Wanyu Wang", "Xian Wu", "Xiangyu Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb64bb5faa37e68b35a2398ad896fc02181a4d2749a3067391e6b3bf023a0f48", "submitted_at": "2026-08-04T07:47:37Z", "updated_at": "2026-08-04T07:47:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.03270v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.03270", "identifiers": {"arxiv": "2608.03270"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.03270"}], "name": "GUI-Lens: Coarse-to-Fine Cropping for GUI Grounding with General-Purpose VLMs", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Lens uses OCR and detected interface components as coordinate references while a general-purpose VLM selects progressively enlarged screenshot crops. Instruction-conditioned checks constrain the crop and click proposals before mapping the final location back to the original screen.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2608.03270"}
{"aliases": [], "authors": ["Siqi Fan", "Minghao Li", "Xiaoqian Ma", "Wenhui Tan", "Xiusheng Huang", "Juntong Wu", "Liujie Zhang", "Shuo Shang", "Weihang Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c8692712194d0ff822091d13324e6bb540339103a824f1030f7a971699d0a0b", "submitted_at": "2026-08-04T08:35:51Z", "updated_at": "2026-08-06T07:26:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.03327v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.03327", "identifiers": {"arxiv": "2608.03327"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.03327"}], "name": "Screenshots or Tools? Eliciting Tool Use and Managing Multimodal Context in Hybrid GUI-MCP Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-08-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "This study separates tool-adoption behavior from task competence in hybrid GUI–MCP agents. It tests tool-use incentives and screenshot-history compression, showing why changed action preferences alone need not improve held-out accuracy.", "topics": ["tool-integration", "reinforcement-learning", "efficiency", "interaction-history", "evaluation"], "updated_at": "2026-08-06", "url": "https://arxiv.org/abs/2608.03327"}
{"aliases": [], "authors": ["Boyan Li", "Zhuowen Liang", "Yupeng Xie", "Xiaotian Lin", "Tianqi Luo", "Xinyu Liu", "Yizhang Zhu", "Zhangyang Peng", "Yuan Li", "Zhengxuan Zhang", "Jiayi Zhang", "Nan Tang", "Guoliang Li", "Yuyu Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5efeb3630917f517742e9c45aa422e226342fa3c02c8683ce8bf53a66ad121e6", "submitted_at": "2026-08-04T10:48:11Z", "updated_at": "2026-08-04T10:48:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.03451v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.03451", "identifiers": {"arxiv": "2608.03451"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.03451"}], "name": "DataSpace: Benchmarking Data Agents for Verifiable Analytics over Heterogeneous Workspaces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2608.03451"}
{"aliases": [], "authors": ["Prince Zizhuang Wang", "Aojie Yuan", "Haiyue Zhang", "Xiyang Hu", "Yue Zhao", "Shuli Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6ae30fa3df1d3a96351e33d117fd1661bff5e57f67b78ccd8f7921a7de2358ff", "submitted_at": "2026-08-04T11:42:26Z", "updated_at": "2026-08-04T11:42:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.03499v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.03499", "identifiers": {"arxiv": "2608.03499"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.03499"}], "name": "WeClawArena: An Auditable Sandbox and Benchmark for Cross-User Agents Collaboration and Security in Human-Centered Agent Networks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2608.03499"}
{"aliases": [], "authors": ["Yu Ran", "Wentao Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "152e21cd4f1f290aafadfdcfaa10e15a03378404540a46b5f1a5464a2d432d23", "submitted_at": "2026-08-04T14:34:41Z", "updated_at": "2026-09-28T13:07:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.03740v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.03740", "identifiers": {"arxiv": "2608.03740"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.03740"}], "name": "MissClick: Execution-Aware Adversarial Attacks on Coordinate Generation in GUI Grounding Models", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "MissClick designs white-box attacks around the numerical effect of coordinate digits on executable GUI clicks. Separate untargeted displacement and targeted place-weighted objectives are evaluated across desktop, web and mobile grounding settings.", "topics": ["safety-security", "grounding", "robustness", "evaluation"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2608.03740"}
{"aliases": [], "authors": ["Zhuoran Zhang", "Bowen Li", "Jingcheng Ju", "Yang Shi", "Qixun Wang", "Haotian Wang", "Wei Chen", "Tengjiao Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9315b92fb8fdf4b99b984d9601ddcc3d9a75357d74660644324211f2e02eb54d", "submitted_at": "2026-08-05T07:02:30Z", "updated_at": "2026-08-05T07:02:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.04530v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2608.04530", "identifiers": {"arxiv": "2608.04530"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.04530"}], "name": "FocusMem: Factorizing Content, Readout, and Trust in Latent GUI Memory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-05", "url": "https://arxiv.org/abs/2608.04530"}
{"aliases": [], "authors": ["Benedikt Schaible", "Anirudh Suresh Bharadwaj", "Ulf Schlichtmann", "Jiang Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e3ffc861e86469082ef0c5ccabfc5951249ce09a9deb6ca3ec17cbe8381b41d", "submitted_at": "2026-08-05T11:58:02Z", "updated_at": "2026-08-05T11:58:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.04734v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.04734", "identifiers": {"arxiv": "2608.04734"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.04734"}], "name": "A Systolic Array Architecture for Nonlinear Activation Functions and Softmax Computation using Chebyshev Polynomials", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-05", "url": "https://arxiv.org/abs/2608.04734"}
{"aliases": [], "authors": ["Longtao Guo", "Zelin Zhang", "Kaifeng Huang", "Yang Shi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "30e5ce3d86881041491ae507839b2f3d302691aea1d2fdfc92cfc00f35efefe7", "submitted_at": "2026-08-05T12:08:16Z", "updated_at": "2026-08-05T12:08:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.04741v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.04741", "identifiers": {"arxiv": "2608.04741"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.04741"}], "name": "LoginTrap: Uncovering Task-Agnostic Phishing-Style Indirect Prompt Injection Attacks against LLM-based Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces LoginTrap, a black-box indirect-injection evaluation in which attacker-controlled page context and login flows induce unauthorized authentication. It measures end-to-end private-data leakage across web-agent backbones, architectures and defenses.", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2026-08-05", "url": "https://arxiv.org/abs/2608.04741"}
{"aliases": [], "authors": ["Dongsheng Chen", "Yuxuan Li", "Guanhua Chen", "Jiaxin Zhang", "Xiangyu Zhao", "Lei Ma", "Xin Yao", "Xuetao Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f8b84135f72558eea1fae428df7cb30f62232631de49e690e73010aeca22d4c", "submitted_at": "2026-08-05T12:23:16Z", "updated_at": "2026-08-05T12:23:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.04755v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.04755", "identifiers": {"arxiv": "2608.04755"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.04755"}], "name": "\"Allow\" to Achieve, Over-Privileged Inadvertently: The Unintended Cost of Task-Completion-Driven Pop-up Decisions in Mobile GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-08-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "This permission-literacy study tests whether mobile GUI agents grant only task-necessary permissions. Controlled changes to task context and requester identity expose authorization biases and the trade-off between suppressing unnecessary and legitimate grants.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-08-05", "url": "https://arxiv.org/abs/2608.04755"}
{"aliases": [], "authors": ["Jingsheng Zheng", "Xinyuan Fang", "Jintian Zhang", "Zhengke Gui", "Huajun Chen", "Ningyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2ff2c8b60ec9a5ed0639bec49e8ef4c4017e7893c6f48121d8eba95283054b9", "submitted_at": "2026-08-04T17:55:41Z", "updated_at": "2026-08-04T17:55:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.05013v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.05013", "identifiers": {"arxiv": "2608.05013"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.05013"}], "name": "OneDayAgent: Towards a Long-Horizon Harness for Autonomous Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-04", "url": "https://arxiv.org/abs/2608.05013"}
{"aliases": [], "authors": ["Buzhao Liu", "Xinhang Ma", "Yevgeniy Vorobeychik"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "628642253457274941d749876591165fedbdb2ebcc0da76e7c842b1d1d00f957", "submitted_at": "2026-08-05T21:44:44Z", "updated_at": "2026-08-05T21:44:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.05430v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.05430", "identifiers": {"arxiv": "2608.05430"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.05430"}], "name": "Robust Context-Aware Detection of Malicious Instructions in Text", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-05", "url": "https://arxiv.org/abs/2608.05430"}
{"aliases": [], "authors": ["Linqiang Guo", "Wei Liu", "Li Gu", "Yang Wang", "Tse-Hsun", "Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc7934927c82a43e6ae6a246ade1d50c4551cc49d263ca67cfb32906cbec9ea4", "submitted_at": "2026-08-06T04:17:27Z", "updated_at": "2026-08-06T04:17:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.05587v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.05587", "identifiers": {"arxiv": "2608.05587"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.05587"}], "name": "StepReflect: Structured UI Transition Reflection for Mobile GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "StepReflect predicts structured per-action reflections from transition specifications and paired visual evidence. The paper distinguishes offline transition-label accuracy from online task success when the reflection model is integrated into mobile agents.", "topics": ["self-reflection", "state-tracking", "imitation-learning", "reward-models", "evaluation", "efficiency"], "updated_at": "2026-08-06", "url": "https://arxiv.org/abs/2608.05587"}
{"aliases": [], "authors": ["Nossa Iyamu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e5a48c8436a16957f1bda4f45cd412af32a39d075774a2d28cb8fb98d418664d", "submitted_at": "2026-08-06T09:17:43Z", "updated_at": "2026-08-06T09:17:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.05784v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2608.05784", "identifiers": {"arxiv": "2608.05784"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.05784"}], "name": "Activity Frames: Deterministic Screen-Activity Compilation for Agent Memory and Replay", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-06", "url": "https://arxiv.org/abs/2608.05784"}
{"aliases": [], "authors": ["Weikai Xu", "Yunren Feng", "Haoxiang Lei", "Kun Huang", "Yuxuan Liu", "Kang Zhao", "Xiaolin Hu", "Shuo Shang", "Bo An"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce37dd8217afbaa07ddea25c0daf78cdbfb449851ca6ab17b7f3ecacd451eb2f", "submitted_at": "2026-08-06T11:15:41Z", "updated_at": "2026-08-06T11:15:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.05891v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.05891", "identifiers": {"arxiv": "2608.05891"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.05891"}], "name": "AppDeltaWorld: Transition-Grounded Delta Code World Model for Mobile GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "AppDeltaWorld predicts executable code updates for the next mobile UI state, using transition-constrained retrieval and rendered visual assets. It evaluates reconstruction fidelity separately from policy training and adaptation using generated environments.", "topics": ["world-models", "data-generation", "imitation-learning", "reinforcement-learning", "mobile-navigation"], "updated_at": "2026-08-06", "url": "https://arxiv.org/abs/2608.05891"}
{"aliases": [], "authors": ["Weiwei Li", "Junzhuo Liu", "Tong Chu", "Hengfu Yu", "Wen Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb90627131765746e3a138ab0862dd68ea963106db4055760d78041c856acc8b", "submitted_at": "2026-08-06T14:10:23Z", "updated_at": "2026-08-06T14:10:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.06065v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.06065", "identifiers": {"arxiv": "2608.06065"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.06065"}], "name": "The Next Screenshot Knows: Gated Hindsight Distillation for Mobile GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Gated Hindsight Distillation gives a teacher access to the next screenshot during training while the student sees only the available trajectory prefix. Distillation is applied when hindsight repairs a student failure, and the resulting policies are evaluated on Android tasks.", "topics": ["imitation-learning", "interaction-history", "mobile-navigation"], "updated_at": "2026-08-06", "url": "https://arxiv.org/abs/2608.06065"}
{"aliases": [], "authors": ["Jiaming Wei", "Zekun Wu", "Adriano Koshiyama", "Maria Perez-Ortiz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9055ec9c70777fc94179983d82bf841bac12332ebed7f49342883c8d842d0fbc", "submitted_at": "2026-08-06T15:37:04Z", "updated_at": "2026-08-06T15:37:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.06171v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.06171", "identifiers": {"arxiv": "2608.06171"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.06171"}], "name": "Routing Is Least Learnable Where It Is Most Valuable: Bounds on Representation Routing for Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Measures the complementarity, cost and rerun variability of six browser observation modes on WebArena and VisualWebArena. Its routing experiments distinguish an oracle cost opportunity from policies that fail to robustly outperform a fixed observation mode.", "topics": ["screen-understanding", "efficiency", "evaluation", "robustness"], "updated_at": "2026-08-06", "url": "https://arxiv.org/abs/2608.06171"}
{"aliases": [], "authors": ["Zhuoxin Zhan", "Akbar Rafiey", "Avery Ma", "Leila Pishdad", "Layla El Asri"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce2f72dbbf1c60354c73869cef490de3417cdc93cff0ce3ad36f32efeda66639", "submitted_at": "2026-08-06T18:14:30Z", "updated_at": "2026-08-06T18:14:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.06477v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.06477", "identifiers": {"arxiv": "2608.06477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.06477"}], "name": "StepJack: Benchmarking Computer-Use Agent Safety Against Multi-Step Indirect Prompt Injection", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-06", "url": "https://arxiv.org/abs/2608.06477"}
{"aliases": [], "authors": ["Jiahao Zhang", "Yifan Zhang", "Yu Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "febf643f0a512a845134c9404ff2da1e2f67928e86304e00e72b2b7e7fc5183f", "submitted_at": "2026-08-07T05:05:05Z", "updated_at": "2026-08-07T05:05:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.06811v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.06811", "identifiers": {"arxiv": "2608.06811"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.06811"}], "name": "Coupling Planning with Episodic Memory in LLM Agents for Software Issue Resolution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-07", "url": "https://arxiv.org/abs/2608.06811"}
{"aliases": [], "authors": ["Fuyao Zhang", "Jiaming Zhang", "Che Wang", "Boyang Chen", "Yurong Hao", "Xiongtao Sun", "Guowei Guan", "Blaise Delattre", "Yang Cao", "Wei Yang Bryan Lim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6478460b6c4e32f7e52d5f6be01dc5995f28394b822ed86f8090827741eddabf", "submitted_at": "2026-08-07T06:39:51Z", "updated_at": "2026-08-07T06:39:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.06862v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.06862", "identifiers": {"arxiv": "2608.06862"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.06862"}], "name": "SynChain: Inducing Computer-Use Agent Systems to Construct Their Own Attack Chains", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-07", "url": "https://arxiv.org/abs/2608.06862"}
{"aliases": [], "authors": ["Xiaoou Liu", "Longchao Da", "Hanyang Chen", "Yuan Ling", "Hua Wei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4d99685a75d2023db5c9ad943e103ec779eebd037fe7a113848d8a31e08fd68b", "submitted_at": "2026-08-07T21:47:43Z", "updated_at": "2026-08-07T21:47:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.07775v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:40.283922+00:00", "role": "primary", "sha256": "ddd5bd966c7430bd9d4fd0269e7d3d46d8839fa5bfa57d601b97b4dc735cfc9a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:40.283922+00:00", "role": "primary", "sha256": "ddd5bd966c7430bd9d4fd0269e7d3d46d8839fa5bfa57d601b97b4dc735cfc9a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:10:44.898646+00:00", "role": "primary", "sha256": "d4afe27efcce2a3b9401f9677304db3cdedfa66f537293bb742bd1abfb16ee59", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2608.07775v1"}], "id": "arxiv:2608.07775", "identifiers": {"arxiv": "2608.07775"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:40.283922+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.07775"}], "name": "AndroidReality: How Far Are Mobile Agents from the Real World?", "notes": "The reviewed protocol normally activates one perturbation type at a time while preserving task semantics and reward. Results do not establish robustness to arbitrary combined deployment disturbances; experiments were not independently reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3.1 Perturbation formulation; 3.2 Injection; 3.3 Evaluation protocol; 4.3 Error analysis); not an entire-paper reading or reproduction claim.", "published_at": "2026-08-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Extends AndroidWorld with controlled state, transition, and action perturbations to test mobile-agent robustness, and studies a training-free introspective recovery mechanism.", "topics": ["evaluation", "robustness", "self-reflection", "mobile-navigation"], "updated_at": "2026-08-07", "url": "https://arxiv.org/abs/2608.07775"}
{"aliases": [], "authors": ["Mansi Uniyal", "Agamdeep Singh", "Ananya Singha", "Priyanshu Gupta", "Mukul Singh", "Gust Verbruggen", "Vu Le", "Sumit Gulwani"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0141914280c394d8f8ec9f321aed755ad7d46855962886150f2def55e26d9fd", "submitted_at": "2026-08-08T02:57:50Z", "updated_at": "2026-08-08T02:57:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.07873v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.07873", "identifiers": {"arxiv": "2608.07873"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.07873"}], "name": "Back to the Future: A workbook time machine for spread sheet creation benchmarks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-08", "url": "https://arxiv.org/abs/2608.07873"}
{"aliases": [], "authors": ["Zejun Xu", "Taiyi Chen", "Jin Li", "Yongtong Gu", "Qi Cheng", "Aixuan Lv", "Shuai Zhu", "Pengfei Zhu", "Kaichen Yang", "Boyu Sun", "Yixian Yang", "Mulong Xie", "Xin Liu", "Dagang Li", "Xiaoteng Ma", "Hongru Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0e469cd7d84080e11df46d826460bea79eae71b8fa5786cacd96aac02b8b0a8", "submitted_at": "2026-08-09T01:08:19Z", "updated_at": "2026-08-09T01:08:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.08392v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.08392", "identifiers": {"arxiv": "2608.08392"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.08392"}], "name": "CAP: A Scalable Benchmark for Evaluating Cross-Site Browser Agents with Complex Actions and Perception", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "CAP builds cross-site browser workflows by decomposing websites into operation and perception requirements and recomposing them into tasks. Its judge-based evaluation diagnoses complex interactions and visual-perception failures across real websites.", "topics": ["evaluation", "web-navigation", "cross-app", "screen-understanding"], "updated_at": "2026-08-09", "url": "https://arxiv.org/abs/2608.08392"}
{"aliases": [], "authors": ["Rahul Deivasigamani", "Sayeda Faatin Alvi", "Derqui Andrea", "Kaushal Punjabi", "Stjepan Picek"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64e4845ce85dd67b7790cf5de357f12deb72b71eb72031298e1064809099f063", "submitted_at": "2026-08-09T22:13:55Z", "updated_at": "2026-08-09T22:13:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.08939v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2608.08939", "identifiers": {"arxiv": "2608.08939"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.08939"}], "name": "Not an A11y: How Android Accessibility Exposes Mobile AI Agents to Indirect Prompt Injection", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Demonstrates that unsanitized Android accessibility tree metadata exposes mobile GUI agents to stealthy indirect prompt injection attacks.", "topics": ["safety-security", "mobile-navigation", "screen-understanding"], "updated_at": "2026-08-09", "url": "https://arxiv.org/abs/2608.08939"}
{"aliases": [], "authors": ["Weijie Liang", "Yuanfeng Song", "Xing Chen", "Caleb Chen Cao", "Sirui Han", "Yike Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "06ed0c8d53085ea151e68889cb2e382e2a793ea312371fdaaae80ce52dc04008", "submitted_at": "2026-08-10T08:24:19Z", "updated_at": "2026-08-10T08:24:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09268v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.09268", "identifiers": {"arxiv": "2608.09268"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.09268"}], "name": "Can Coding Agents Solve Repository-Level Issues with Rendered Code? An Exploratory Study of Visual Representations", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-10", "url": "https://arxiv.org/abs/2608.09268"}
{"aliases": [], "authors": ["Shengcheng Yu", "Yuchen Ling", "Junyang Xing", "Quan Zhou", "Chunrong Fang", "Zhenyu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fdd7c1df0fdb4999d4297a943bdee6f6da55e49c1170d9aaf292c000688a44c4", "submitted_at": "2026-08-10T08:33:19Z", "updated_at": "2026-08-10T08:33:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09278v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.09278", "identifiers": {"arxiv": "2608.09278"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.09278"}], "name": "Software Engineering for and with GUI Agent", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-10", "url": "https://arxiv.org/abs/2608.09278"}
{"aliases": [], "authors": ["Yuke Li", "Xuehan Hou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f05f128cccb514987b7defb52c8f43b45d917e0e0f9570db4e75c4fe23b38533", "submitted_at": "2026-08-10T14:29:07Z", "updated_at": "2026-08-10T14:29:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09654v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.09654", "identifiers": {"arxiv": "2608.09654"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.09654"}], "name": "Hallucination-Free GUI Grounding via Regression-Free Layout-Aware Matching", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-08-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "This framework separates frozen-model instruction expansion from a trained layout-aware matcher that localizes GUI elements without coordinate regression. Binary text/icon supervision supports the matcher, with evaluations on GUI grounding and Mind2Web element-selection tasks.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2026-08-10", "url": "https://arxiv.org/abs/2608.09654"}
{"aliases": [], "authors": ["Renshan Zhang", "Haoyang Meng", "Yixiao He", "Rui Shao", "April Hua Liu", "Liqiang Nie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87e9b3463232f606aea60df7d86626d5aa188e774b0fe515af43e1be49b1c66b", "submitted_at": "2026-08-10T15:23:52Z", "updated_at": "2026-08-10T15:23:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09723v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.09723", "identifiers": {"arxiv": "2608.09723"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.09723"}], "name": "LookAgain: Closed-Loop GUI Grounding with Visually Grounded Reflection", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "LookAgain treats a GUI coordinate as a revisable hypothesis, rendering its marker and a local screenshot patch before confirmation. Reflective supervised trajectories provide a cold start, followed by GRPO using terminal grounding correctness.", "topics": ["grounding", "self-reflection", "reinforcement-learning", "imitation-learning"], "updated_at": "2026-08-10", "url": "https://arxiv.org/abs/2608.09723"}
{"aliases": [], "authors": ["Santosh Patapati"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5ccb0d41798a1a3236e6f9b4361814cccdd40facf9b43bf0bf52932337484dc", "submitted_at": "2026-07-01T16:24:07Z", "updated_at": "2026-07-01T16:24:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09944v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:45.064380+00:00", "role": "primary", "sha256": "f26fba818a44c6a7f541ab3b0341ca8f4798d837a425a2ac9fb0fc47a553b7bc", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:45.064380+00:00", "role": "primary", "sha256": "f26fba818a44c6a7f541ab3b0341ca8f4798d837a425a2ac9fb0fc47a553b7bc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:05:26.923383+00:00", "role": "primary", "sha256": "777197864eeaa34d180bf6b91f32408b7b3d8cc5c9faecc95eca276f5bd426de", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/html/2608.09944v1"}], "id": "arxiv:2608.09944", "identifiers": {"arxiv": "2608.09944"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:45.064380+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.09944"}], "name": "Navigation Alone Is Not Enough: Evaluating Explanatory Assistive UI Agents", "notes": "Packaged tasks use recorded linear transition graphs. Reported model experiments use a text-input validation subset, so their results are not evidence of screenshot-based navigation performance. No results were independently reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3.3 Interface states and task annotations; 3.4 Evaluation; 4.1 Models and experimental setup; 5 Discussion); not an entire-paper reading or reproduction claim.", "published_at": "2026-07-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces NeXUI, an assistive UI-agent benchmark that evaluates navigation together with interface-grounded explanations and user oversight in recorded interface states.", "topics": ["evaluation", "human-in-the-loop", "web-navigation"], "updated_at": "2026-07-01", "url": "https://arxiv.org/abs/2608.09944"}
{"aliases": [], "authors": ["Zelei Cheng", "Amritansh Mishra", "Sambit Sahu", "William Campbell"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4767a9d2ac1aabf0b6aa1e44956fbb6e69b02c56704397e64c7394e92a17cdce", "submitted_at": "2026-08-11T01:30:56Z", "updated_at": "2026-08-11T01:30:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.10357v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.10357", "identifiers": {"arxiv": "2608.10357"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.10357"}], "name": "Efficient Reinforcement Learning for Long-Horizon Tool-Use Agentic Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-11", "url": "https://arxiv.org/abs/2608.10357"}
{"aliases": [], "authors": ["Mizanur Rahman", "Mohammed Saidul Islam", "Ridwan Mahbub", "Md Tahmid Rahman Laskar", "Shafiq Joty", "Enamul Hoque Prince"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1089e57e0f0146f29a4a411c5c6f72da3c9f45be223fbbc5c9630315be6a680", "submitted_at": "2026-08-11T01:45:56Z", "updated_at": "2026-08-11T01:45:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.10366v1"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:26Z", "role": "primary", "sha256": "2158cabc9a6211e0b75e8e3307f80ebc0d42c42fccc03d0143dcf574c2f663db", "supports": ["summary", "environments", "scope"], "url": "https://github.com/vis-nlp/DSAgentBench"}], "id": "arxiv:2608.10366", "identifiers": {"arxiv": "2608.10366"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.10366"}], "name": "DSAgentBench: Can Agents Automate End-to-End Data-Science Workflows in Real Computer Environments?", "notes": "The paper abstract and official project README were reviewed. The README documents an Ubuntu desktop and GUI action interface and includes LLM-assisted evaluators; this annotation does not assert that every evaluator is deterministic. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "DSAgentBench evaluates end-to-end data-science workflows spanning analysis, modeling and visualization inside a real desktop environment. Task evaluators check produced analytical and visual artifacts, while the released interface supports screenshot or accessibility observations and GUI actions.", "topics": ["evaluation", "workflow", "science", "long-horizon"], "updated_at": "2026-08-11", "url": "https://arxiv.org/abs/2608.10366"}
{"aliases": [], "authors": ["Zixing Chen", "Xingyuan Liu", "Jie Zhu", "Huaixia Dou", "Shuo Jiang", "Junhui Li", "Lifan Guo", "Feng Chen", "Chi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ec2768e4591e7307b7bcc15ba28b7905661ca8bdb56dc0eacbf0a223c9ba8654", "submitted_at": "2026-08-11T08:48:54Z", "updated_at": "2026-08-11T08:48:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.10669v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.10669", "identifiers": {"arxiv": "2608.10669"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.10669"}], "name": "REDAgentBench: Executable Red Teaming and Faithful Measurement of LLM Agent Systems", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-11", "url": "https://arxiv.org/abs/2608.10669"}
{"aliases": [], "authors": ["Junjie Ye", "Zhuohui Sheng", "Shaofan Liu", "Yulun Zhu", "Wenjie Fu", "Dingwei Zhu", "Ming Zhang", "Yujiong Shen", "Weichao Wang", "Xin Zhao", "Shihan Dou", "Tao Gui", "Qi Zhang", "Xuanjing Huang", "Pluto Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14ec426c81b302631031dff7b510ff4ffc452e03bc6d8a30b06d8ca350a12aa4", "submitted_at": "2026-08-11T09:14:53Z", "updated_at": "2026-08-11T09:14:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.10692v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2608.10692", "identifiers": {"arxiv": "2608.10692"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.10692"}], "name": "SPIEval: Evaluating Large Language Models as Mobile Assistants over Scattered Personal Information", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-11", "url": "https://arxiv.org/abs/2608.10692"}
{"aliases": [], "authors": ["Albus W. Ng", "Yi Han", "Jusheng Zhang", "Wenhao Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b3819f8acc2510dc86b6522340ea8b4d5305202568484ac2515a3f024e04dea", "submitted_at": "2026-08-11T08:01:05Z", "updated_at": "2026-08-11T08:01:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.11274v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.11274", "identifiers": {"arxiv": "2608.11274"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.11274"}], "name": "Agent Safety Should Be a Runtime Contract", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-11", "url": "https://arxiv.org/abs/2608.11274"}
{"aliases": [], "authors": ["Saptarshi Biswas", "James I. Lathrop", "Rana D. Parshad"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1344dc28e0edaeec2a35bc7493c276bc9c830bfaf9e113339e0e7f8dfb4e2ec0", "submitted_at": "2026-08-11T19:15:18Z", "updated_at": "2026-09-07T03:07:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.11362v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.11362", "identifiers": {"arxiv": "2608.11362"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.11362"}], "name": "RevCRN: Reversible Analog Computation using Chemical Reaction Networks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-07", "url": "https://arxiv.org/abs/2608.11362"}
{"aliases": [], "authors": ["Ziqiang Wang", "Li Gu", "Zhixiang Chi", "Zhi Liu", "Seyed Mehdi Ayyoubzadeh", "Yuanhao Yu", "Yang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2606ae96ca3ea99b5c7278485601f9bde3b64eefdd3bfecc2bbfc678c39b4fc", "submitted_at": "2026-08-11T21:00:46Z", "updated_at": "2026-08-11T21:00:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.11434v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.11434", "identifiers": {"arxiv": "2608.11434"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.11434"}], "name": "Benchmarking LLM Judges for Mobile Agent Evaluation", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-08-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileJudgeBench compares completion judges against human labels on mobile-agent trajectories. It studies screenshot-sampling baselines, model-dependent error profiles and how judge quality relates to agent rankings and reinforcement-learning rewards.", "topics": ["evaluation", "reward-models", "robustness"], "updated_at": "2026-08-11", "url": "https://arxiv.org/abs/2608.11434"}
{"aliases": [], "authors": ["Linqiang Guo", "Li Gu", "Zihuan Jiang", "Zhixiang Chi", "Siobhan Reid", "Ziqiang Wang", "Yuanhao Yu", "Wei Liu", "Yang Wang", "Tse-Hsun", "Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f7e00db81600eb5f946dae4a260ffd174ca4ceb3890f73168714547b298fae86", "submitted_at": "2026-08-12T02:55:50Z", "updated_at": "2026-08-12T02:55:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.11588v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2608.11588", "identifiers": {"arxiv": "2608.11588"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.11588"}], "name": "CoAdapt-GUI: Joint Workflow Context and Policy Adaptation for Unseen GUI Applications", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-12", "url": "https://arxiv.org/abs/2608.11588"}
{"aliases": [], "authors": ["Ruitao Wang", "Yuwen Hao", "Menglin Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf543ef0b77eb6fe430c8cef5382932f13e4e4136aaba9c14c44b5a591d6612a", "submitted_at": "2026-08-12T12:03:20Z", "updated_at": "2026-08-12T12:03:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.12429v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.12429", "identifiers": {"arxiv": "2608.12429"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.12429"}], "name": "SynWeaver: Website-Prior Task and Trajectory Co-Synthesis for Web Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-08-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "SynWeaver explores a website to build a map of page states and executable interactions, then learns website priors for task proposals. Joint task/trajectory synthesis revises inconsistent pairs and verifies or repairs the resulting supervision.", "topics": ["data-generation", "web-navigation", "state-tracking", "robustness"], "updated_at": "2026-08-12", "url": "https://arxiv.org/abs/2608.12429"}
{"aliases": [], "authors": ["Churchil Dwivedi", "Hiya Shah", "Hemanga Tahbildar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f627ba3336ac86d95066596b7ad6d1dace3d8cf92c4aba5993538c7c932d9569", "submitted_at": "2026-08-14T06:16:33Z", "updated_at": "2026-09-30T14:15:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.13991v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.13991", "identifiers": {"arxiv": "2608.13991"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.13991"}], "name": "On the Acceleration of Pulsar Timing computations using Normalising Flows and Parallelisation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2608.13991"}
{"aliases": [], "authors": ["Thilo Reintjes", "Sivajeet Chand", "Derui Zhu", "Sushant Kumar Pandey", "Alexander Pretschner"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8a6bd69d421bdfb9a21c22e715a59d0afb0b0cab75510f28ff569903ecffae5", "submitted_at": "2026-08-14T09:36:33Z", "updated_at": "2026-08-14T09:36:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.14131v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.14131", "identifiers": {"arxiv": "2608.14131"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.14131"}], "name": "LegacyWorld: Atomicity-Aware Evaluation of GUI Agents for Legacy Workflows", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "LegacyWorld evaluates Windows GUI workflows under an atomicity criterion: finish correctly or fail without persistent unintended changes. It compares expert prompts with prompts derived from recorded golden paths and separates useful completion, safe failure and non-atomic side effects.", "topics": ["evaluation", "safety-security", "workflow", "recording"], "updated_at": "2026-08-14", "url": "https://arxiv.org/abs/2608.14131"}
{"aliases": [], "authors": ["Xiaokai Yan", "Jingtao Ding", "Yong Li", "Zhiwen Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e2660e01921a78bed7280b34f35a1b587e62339c2e280d6294bbb6d6e666786e", "submitted_at": "2026-08-14T09:36:59Z", "updated_at": "2026-08-14T09:36:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.14132v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:40.283922+00:00", "role": "primary", "sha256": "ddd5bd966c7430bd9d4fd0269e7d3d46d8839fa5bfa57d601b97b4dc735cfc9a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:40.283922+00:00", "role": "primary", "sha256": "ddd5bd966c7430bd9d4fd0269e7d3d46d8839fa5bfa57d601b97b4dc735cfc9a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:05:22.211464+00:00", "role": "primary", "sha256": "8d6aa8484ff4a7cbaac589c2c220407ebf6ecd1e216015a248125abdb0f2ed2d", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/html/2608.14132v1"}], "id": "arxiv:2608.14132", "identifiers": {"arxiv": "2608.14132"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:40.283922+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.14132"}], "name": "Act2Intention: A Benchmark For Developing Active Mobile Agents Through Inferring User Intention from GUI Actions", "notes": "The benchmark combines real and generated trajectory material. Opportune intervention timing and real-user acceptance are stated limitations. Section 4.2.2 specifies Android emulator instances for generated action trajectories. No results were independently reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3 Task definition; 4.1 Raw data collection; 4.2 Trajectory construction; 7 Limitations and future directions); not an entire-paper reading or reproduction claim.", "published_at": "2026-08-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces intention understanding, prediction, and confirmed execution tasks over mobile GUI trajectories, with a proactive agent combining personalized intention models and experience-guided actions.", "topics": ["human-in-the-loop", "planning", "mobile-navigation", "evaluation"], "updated_at": "2026-08-14", "url": "https://arxiv.org/abs/2608.14132"}
{"aliases": [], "authors": ["Joshua Adamek", "Natalie Carl", "Trever Schirmer", "Moritz Heinlein", "David Bermbach", "Sergio Lucia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f02e6b123dd26e8ed15ce9d8ed1acd615e2249b4e254453c5e6a26f15929a78d", "submitted_at": "2026-08-14T16:08:49Z", "updated_at": "2026-08-26T13:00:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.14427v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.14427", "identifiers": {"arxiv": "2608.14427"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.14427"}], "name": "Large-scale workflow placement in serverless computing using integer nonlinear programming", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-26", "url": "https://arxiv.org/abs/2608.14427"}
{"aliases": [], "authors": ["Ziheng Qin", "Yaxin Lu", "Zhangyang Atlas Wang", "Kai Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cabe4f33d653ccaa482bba0d13f5ac2a577d7f83a02fd8ac76ece228b0427328", "submitted_at": "2026-08-15T07:16:28Z", "updated_at": "2026-08-15T07:16:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.15089v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "id": "arxiv:2608.15089", "identifiers": {"arxiv": "2608.15089"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.15089"}], "name": "StateM: Reaching 95.3% Raw Accuracy, or a \\$15 Frontier Run, on Terminal-Bench 2.1 via Harness Scaling", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-15", "url": "https://arxiv.org/abs/2608.15089"}
{"aliases": [], "authors": ["Zihan Ding", "Longxu Dou", "Qi Gao", "Xiangwu Guo", "Shengchao Hu", "Zilong Huang", "Zihang Jiang", "Lei Ke", "Mengcheng Lan", "Weixian Lei", "Hanxuan Li", "Honglin Li", "Xiyun Li", "Zaitang Li", "Leowei Liang", "Xin Luo", "Haozhe Ma", "Jiayi Mao", "Zhoujie Pan", "Can Qin", "Tianyuan Qu", "Weiqi Wang", "Wenkai Wang", "Yonglin Wang", "Yuxin Wang", "Chenxu Wu", "Yingchen Yu", "Chenyu Zhang", "Yuhao Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "efe635792c696bca73c9a436be67d8528582fb239d4d4f8d3de531219b1f4f8e", "submitted_at": "2026-08-16T20:59:59Z", "updated_at": "2026-08-16T20:59:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.15930v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.15930", "identifiers": {"arxiv": "2608.15930"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.15930"}], "name": "UI-Mate: Advancing Open-Weight Foundation GUI Agents with In-Context Demonstrations", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines environment-grounded training with demonstration-conditioned workflows, and introduces OSWorkerBench for extended office-task evaluation.", "topics": ["workflow", "data-generation", "long-horizon"], "updated_at": "2026-08-16", "url": "https://arxiv.org/abs/2608.15930"}
{"aliases": [], "authors": ["Junqi Liu", "Yufan He", "Yexiao He", "Pengfei Guo", "Dong Yang", "Andriy Myronenko", "Can Zhao", "Hanrong Ye", "Tianhao Qi", "Yuyin Zhou", "Daguang Xu", "Yucheng Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a42e676ca2420ea1870ab6944bfcf35ebce98ecc12630e5f6fe186a220652932", "submitted_at": "2026-08-17T07:44:57Z", "updated_at": "2026-08-17T07:44:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.16211v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.16211", "identifiers": {"arxiv": "2608.16211"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.16211"}], "name": "BaT: Towards Self-Evolving Medical Research Agent with Stage Rubrics", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-17", "url": "https://arxiv.org/abs/2608.16211"}
{"aliases": [], "authors": ["Xiangfan Wu", "Zonghao Ying", "Huiyu Wu", "Xing Zheng", "Huangsheng Cheng", "Xiaorong Shi", "Jing Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8df04718c5fa3eb910dc8163ffbbd299ea1c45881671152beee81f30bb2c2b1e", "submitted_at": "2026-08-17T10:41:18Z", "updated_at": "2026-08-17T10:41:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.16391v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.16391", "identifiers": {"arxiv": "2608.16391"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.16391"}], "name": "Ventor-QTest: Threat-Model-Driven Verification of Vendor-Hosted LLM APIs", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-17", "url": "https://arxiv.org/abs/2608.16391"}
{"aliases": [], "authors": ["Tong Zhang", "Motasem Alfarra", "Carlos Hinojosa", "Christos Louizos", "Bernard Ghanem"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51dc1b5b6207d0115ba3960a7c1eee251b6c708c2e583810e67988c5c079eeb0", "submitted_at": "2026-08-17T19:13:55Z", "updated_at": "2026-08-17T19:13:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.17067v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.17067", "identifiers": {"arxiv": "2608.17067"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.17067"}], "name": "DiSCO: Defending text-to-image generation through distribution-guided contrastive prompt optimization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-17", "url": "https://arxiv.org/abs/2608.17067"}
{"aliases": [], "authors": ["Zhi Zheng", "Rongsheng Chen", "Yunpeng Ba", "Zhenkun Wang", "Yee Whye Teh", "Wee Sun Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0404b10d82c520c75afa8ad517b2292771b48d019c186e0dff7753672a8703d", "submitted_at": "2026-08-18T03:03:53Z", "updated_at": "2026-08-21T05:30:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.17310v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "id": "arxiv:2608.17310", "identifiers": {"arxiv": "2608.17310"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.17310"}], "name": "Agentic ESOpt: Fine-Tuning Long-Horizon LLM Agents with Minimal GPU Requirements", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-21", "url": "https://arxiv.org/abs/2608.17310"}
{"aliases": [], "authors": ["AIMAE Team", "Tianxiang Chen", "Yan Cheng", "Zhangye Han", "Xiaowei Li", "Chang Liu", "Cheng Liu", "Zhongqiang Ma", "Long Peng", "Xiaobing Tu", "Yinggui Wang", "Hongliang Wei", "Chen Wu", "Daiping Xin", "Kunyu Zhou", "Pengyang Zhou", "Peiyuan Chen", "Ziyuan Chen", "Yutao Deng", "Chunyu Dong", "Xiangyu Fu", "Yicheng Feng", "Ruian He", "Haochen Li", "Miancan Liu", "Zhengqin Liu", "Wei Peng", "Jinkui Ren", "Haoyu Tan", "Dong Xiao", "Rongkun Xue", "Shujian Yang", "Xianhang Ye", "Ziqi Yuan", "Ziyang Yu", "Linghan Zhang", "Xiantao Zhang", "Xuanpu Zhao", "Yinan Zhao", "Zhenghui Zhao", "Bin Zhu", "Likai Zou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "474187aaa79165dc8652eec1e2af6b1733b2ee37b3cd125c89ff0defcaf2e193", "submitted_at": "2026-08-18T03:23:31Z", "updated_at": "2026-08-18T03:23:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.17319v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.17319", "identifiers": {"arxiv": "2608.17319"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.17319"}], "name": "Wuying-Browser-Agent: Real-World Centric Fundamental Long-Horizon Browser Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines a browser execution framework, recovery-oriented supervision, and online reinforcement learning, with a bilingual live-web task benchmark.", "topics": ["web-navigation", "long-horizon", "reinforcement-learning"], "updated_at": "2026-08-18", "url": "https://arxiv.org/abs/2608.17319"}
{"aliases": [], "authors": ["Sujin Chen", "Lijun Li", "Tianyi Du", "Jing Shao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c49a67b38314152b83e793b9ade7ce4102a44aa27c331504a8758b94df906c67", "submitted_at": "2026-08-18T11:33:22Z", "updated_at": "2026-08-18T11:33:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.17659v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.17659", "identifiers": {"arxiv": "2608.17659"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.17659"}], "name": "MobileWorldSafety: Benchmarking GUI Agent Safety Against Environmental Injection Attacks in Android Apps", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates environmental injection risks in Android tasks using programmatic risk checks and model-based adjudication of ambiguous cases.", "topics": ["safety-security", "evaluation"], "updated_at": "2026-08-18", "url": "https://arxiv.org/abs/2608.17659"}
{"aliases": [], "authors": ["Rui-Huan Wang", "Si-Tong Wei", "Jia-Qi He", "Heng-Yi Wei", "Baoquan Chen", "Peng-Shuai Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33712da953c3b03e226d16886c54a012699777fce2d5d4049db0d8015e309c6a", "submitted_at": "2026-08-18T16:27:44Z", "updated_at": "2026-08-18T16:27:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.17975v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.17975", "identifiers": {"arxiv": "2608.17975"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.17975"}], "name": "aDSL: Agentic 3D Creation via Joint Agent-Program Design", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-18", "url": "https://arxiv.org/abs/2608.17975"}
{"aliases": [], "authors": ["Tianchen Guan", "Xinlei Lin", "Royce Cheng-Yue", "Xiangjun Wang", "Shuyan Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de4889c021f35c169092e17801fec414cce21b37c74a4ff72af6a6c67a1a0a25", "submitted_at": "2026-08-18T20:38:26Z", "updated_at": "2026-08-18T20:38:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.18307v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.18307", "identifiers": {"arxiv": "2608.18307"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.18307"}], "name": "ComponentBench: Diagnosing Component-Level Failures in Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-08-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "ComponentBench evaluates short interactions with canonical web UI components using verified tasks and matched human reference trajectories. Its diagnostic pipeline compares observation/action representations and analyzes structural difficulty, failure patterns and interaction efficiency.", "topics": ["evaluation", "screen-understanding", "efficiency", "robustness"], "updated_at": "2026-08-18", "url": "https://arxiv.org/abs/2608.18307"}
{"aliases": [], "authors": ["Zhuochun Li", "Youngmin Ko", "Ali Keramati", "Tuhin Kundu", "Liang-Chun Tsai", "Nicola Ferri", "Mirco Milletari", "Jiaxiang Liu", "Susana Palmaz Lopez Pelaez", "Yuepeng Wang", "Vadim Smolyakov", "Xiang Jiang", "Kjartan Olafsson", "Tommy Guy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cbc941d961a6d26823030e3a8ed0e18756dbf526667043377881527b262b53d3", "submitted_at": "2026-08-20T07:37:57Z", "updated_at": "2026-10-01T16:14:54Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2608.19741v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.19741", "identifiers": {"arxiv": "2608.19741"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.19741"}], "name": "One Success Isn't Reliability: Thinkingbox, a Sandbox and Benchmark for Agents in Stateful Business Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2608.19741"}
{"aliases": [], "authors": ["Yucheng Jiang", "Zora Zhiruo Wang", "Ruishi Chen", "Diyi Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a85b3a99d551fcdffdd306d839b39cdfbe7f0f7654aa9adbf81f939aa874e3b9", "submitted_at": "2026-08-20T17:57:00Z", "updated_at": "2026-08-20T17:57:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.20319v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.20319", "identifiers": {"arxiv": "2608.20319"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.20319"}], "name": "Inducing Task Models from Computer-Use Traces", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Task Model Induction disentangles concurrent activities in screenshot and mouse/keyboard traces, then constructs hierarchical goals and procedural control flow for each task. Evaluation tests reconstruction of observed execution and reuse of derived skills on held-out tasks.", "topics": ["recording", "workflow", "memory", "interaction-history"], "updated_at": "2026-08-20", "url": "https://arxiv.org/abs/2608.20319"}
{"aliases": [], "authors": ["Jiakai Tang", "Yan Mi", "Jing Yu", "Yang Zhang", "See-Kiong Ng", "Qi Cao", "Fei Sun", "Xu Chen", "Wen Chen", "Jian Wu", "Han Zhu", "Bo Zheng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52e89b35f989f17fce85c418e748f1de2d7bde635c078c7e7de29444d10ac2b8", "submitted_at": "2026-08-21T03:24:32Z", "updated_at": "2026-08-21T03:24:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.20707v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.20707", "identifiers": {"arxiv": "2608.20707"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.20707"}], "name": "Towards Faithful Simulation of Human Shopping Behavior", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-21", "url": "https://arxiv.org/abs/2608.20707"}
{"aliases": [], "authors": ["Pengshuai Yang", "Zijing Gao", "Xue Yu", "Benhui Zhuang", "Bo Yuan", "Junlan Feng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "28c1d4284da8ba18bf3d40a0402a9ec4d6b85cdb381e4ce9da3ae675c1cc741f", "submitted_at": "2026-08-21T07:16:36Z", "updated_at": "2026-08-21T07:16:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.20797v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.20797", "identifiers": {"arxiv": "2608.20797"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.20797"}], "name": "Automated Trajectory Evaluation for Mobile Agents via Step-Level Consequence Reasoning and Aggregation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-21", "url": "https://arxiv.org/abs/2608.20797"}
{"aliases": [], "authors": ["Yibo Peng", "Long Lian", "David Wagner", "Sizhe Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6678ea148a9cd48f3c775bd55ea12ea7ea7cc62a5379c5c509ff4bc49895a4f7", "submitted_at": "2026-08-21T16:14:07Z", "updated_at": "2026-08-21T16:14:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.21500v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.21500", "identifiers": {"arxiv": "2608.21500"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.21500"}], "name": "SecOPD: Mitigating Adaptive Prompt Injections by On-Policy Distillation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-21", "url": "https://arxiv.org/abs/2608.21500"}
{"aliases": [], "authors": ["Qijia Chen", "Giulio Jacucci"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cabd094270d01b20df7b029be29c5d460d10beb84e43e3e783f4f11997b74114", "submitted_at": "2026-08-22T06:30:03Z", "updated_at": "2026-08-22T06:30:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.21794v1"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.21794", "identifiers": {"arxiv": "2608.21794"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.21794"}], "name": "Lexical Coupling in GUI Element Grounding: Sentence Embeddings Track Labels across Mobile and Web", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies whether text-embedding GUI grounding scores reflect semantic matching or recovery of visible labels. Same-screen ranking comparisons with lexical baselines expose differences between deployable fusion, oracle selection, and performance on label-poor targets.", "topics": ["grounding", "evaluation", "screen-understanding"], "updated_at": "2026-08-22", "url": "https://arxiv.org/abs/2608.21794"}
{"aliases": [], "authors": ["Chengyang Gu", "Le Zhang", "Jingbo Zhou", "Yize Chen", "Yu Shi", "Siqi Bao", "Zheng-Fan Wu", "Hua Wu", "Hui Xiong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9434f8da41679177f6978c039bd42fcacce543de2bd0e4ce05b62ab4d61ed761", "submitted_at": "2026-08-22T08:01:47Z", "updated_at": "2026-08-22T08:01:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.21830v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.21830", "identifiers": {"arxiv": "2608.21830"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.21830"}], "name": "Beyond Success and Failure: Length-Aware Contrastive Learning for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces LACL-GUI, a contrastive learning objective that distinguishes trajectory quality within both successful and failed GUI executions. It favors concise successes and uses divergence from successful trajectories to differentiate failures during policy optimization.", "topics": ["reinforcement-learning", "efficiency", "action-prediction"], "updated_at": "2026-08-22", "url": "https://arxiv.org/abs/2608.21830"}
{"aliases": [], "authors": ["Md Abrar Jahin", "Md Rizwan Parvez"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "970441704314fabffc46cb4a340dd6e0bde0e84ab9b7fcd2db2c9a8be9dc61d8", "submitted_at": "2026-08-22T08:04:59Z", "updated_at": "2026-08-27T07:03:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.21832v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.21832", "identifiers": {"arxiv": "2608.21832"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.21832"}], "name": "GUI-Primitives: Diagnosing Spatial Reasoning Failures in Vision-Language GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI-Primitives, a paired-instruction benchmark that changes a spatial relation while holding the screenshot and anchor fixed. Its diagnostics separate failure to localize either candidate from failure to choose the candidate satisfying the relation.", "topics": ["grounding", "evaluation", "screen-understanding"], "updated_at": "2026-08-27", "url": "https://arxiv.org/abs/2608.21832"}
{"aliases": [], "authors": ["Chenghao Zhang", "Yuxi Cheng", "Saisai Hu", "Canran Xiao", "Yikai Mao", "Dan Roth"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d862afa969c0b4d1c6edc8d126017749844158f786b29c3b0710afecdec2008", "submitted_at": "2026-08-22T10:18:45Z", "updated_at": "2026-09-28T15:54:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.21898v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.21898", "identifiers": {"arxiv": "2608.21898"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.21898"}], "name": "Training Needs Trustworthy Worlds: Verified Synthetic Web Environments for Agent Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds synthetic websites with verified navigation, backend state, and feasible tasks to provide reliable supervision for web-agent training.", "topics": ["data-generation", "sandbox", "reinforcement-learning"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2608.21898"}
{"aliases": [], "authors": ["Mian Zhang", "Manasi Sharma", "Sheng Zhang", "Minglai Yang", "Kejian Shi", "Ying Liu", "Zhiyu Zoey Chen", "Daniel Yue Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2dcd0aff503b37a98c15cfa4dea6ddbd0935eaabca69cfded8289daa9aca12d4", "submitted_at": "2026-08-22T18:48:37Z", "updated_at": "2026-08-22T18:48:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.22077v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.22077", "identifiers": {"arxiv": "2608.22077"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.22077"}], "name": "Spine-Branch Coordination for Multi-agent Computer Use", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-22", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-22", "url": "https://arxiv.org/abs/2608.22077"}
{"aliases": [], "authors": ["YuanHang Xiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb629137ba8c90bc5a773911fef3a162a72123452a4b164039a343d7e31c53f4", "submitted_at": "2026-08-23T17:09:02Z", "updated_at": "2026-08-23T17:09:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.22510v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.22510", "identifiers": {"arxiv": "2608.22510"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.22510"}], "name": "ClawProBench: Trace-Aware Evaluation of AI Agents with Runtime Coverage and Frozen Workplace-Style Holdouts", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-23", "url": "https://arxiv.org/abs/2608.22510"}
{"aliases": [], "authors": ["Zheyuan Deng", "Binghang Lu", "Hanqi Feng", "Shirley Huang", "Dianzhuo Wang", "Yuanda Xu", "Zhiwei Zhang", "Yige Sun", "Changhong Mou", "Runyu Zhang", "Yuexing Hao", "Barnabas Poczos", "Guang Lin", "Xiaomin Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b3a427c6e49c38bbb6871d26358db68976c09fae380a6faa0b3b2425fd169e21", "submitted_at": "2026-08-23T18:11:26Z", "updated_at": "2026-09-30T05:36:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.22533v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.22533", "identifiers": {"arxiv": "2608.22533"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.22533"}], "name": "CONTRAMEM: Learning Self-Evolving Procedural Memory from Contrasting Multi-Model Trajectories", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2608.22533"}
{"aliases": [], "authors": ["Jiaxuan Luo", "Zhanfeng Liao", "Jiayao Teng", "Yuan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ea33a82ab7543a64b60ee3024c6f3e39f88b4e3f06eff94f8c790c56cf76715", "submitted_at": "2026-08-23T20:04:21Z", "updated_at": "2026-08-25T04:24:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.22577v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.22577", "identifiers": {"arxiv": "2608.22577"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.22577"}], "name": "CausalCache: Conditional High-Fidelity Restoration for Long-Horizon GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CausalCache, which selects archived screenshots to restore alongside text summaries under a fixed visual-memory budget. A history-gated adapter processes restored images, with evaluation distinguishing standard step limits from longer-horizon and mobile-transfer settings.", "topics": ["memory", "long-horizon", "efficiency"], "updated_at": "2026-08-25", "url": "https://arxiv.org/abs/2608.22577"}
{"aliases": [], "authors": ["Long Zhang", "Yuhan Chen", "Chaoran Zhang", "Wanxia Cao", "Kun Huang", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Chenliang Li", "Lixin Zou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c1198ce65d5a2fcffc51b37bc4553b35b9e0d809757dc4181747c57c7ae443d", "submitted_at": "2026-08-24T06:27:47Z", "updated_at": "2026-08-24T06:27:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.22847v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.22847", "identifiers": {"arxiv": "2608.22847"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.22847"}], "name": "GSAR: Goal-State-Anchor Rewards for Mobile GUI Agents with Self-Evolving Data Synthesis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GSAR, which synthesizes GUI tasks and successful goal states, then annotates task-relevant interface elements as reward anchors. These anchors support offline trajectory verification and online policy training.", "topics": ["data-generation", "reward-models", "reinforcement-learning"], "updated_at": "2026-08-24", "url": "https://arxiv.org/abs/2608.22847"}
{"aliases": [], "authors": ["Yi Zhu", "Xiongwei Wu", "Qiyi Wang", "Tingyu Qu", "Jiajun Liu", "Sihan Cao", "Long Chen", "Weigao Sun", "Feida Zhu", "Yiran Zhong", "Steven Hoi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d1511aebba0d9714f347a404953fe46211697db7cc092c3f0cd1a6e847c1f6f4", "submitted_at": "2026-08-24T09:38:24Z", "updated_at": "2026-08-25T04:26:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.23035v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.23035", "identifiers": {"arxiv": "2608.23035"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.23035"}], "name": "MobilePA-Bench: Benchmarking Mobile Planner Agents on Complex Real-World Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-24", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-25", "url": "https://arxiv.org/abs/2608.23035"}
{"aliases": [], "authors": ["Yijie Ma", "Chaoyue Niu", "Fan Wu", "Guihai Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7e4fe2cb74124b2897d86a4fee32113164b86a8f5abae040bf77db4900e0bc4", "submitted_at": "2026-08-25T03:10:01Z", "updated_at": "2026-08-25T03:10:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24015v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.24015", "identifiers": {"arxiv": "2608.24015"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24015"}], "name": "Reflection with Action-Induced Visual Differences for Desktop GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Evidence-First Reflection, which separates extraction of action-induced screen changes from judgment of whether an operation achieved its objective. It filters visual differences into relevant evidence before a desktop agent’s reflector decides how execution should continue.", "topics": ["self-reflection", "screen-understanding", "robustness"], "updated_at": "2026-08-25", "url": "https://arxiv.org/abs/2608.24015"}
{"aliases": [], "authors": ["Lin-Fa Lee", "YI-YU Chang", "Kuo-Hui Yeh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bddfe5b5b2312e9b71d01dbb6bb23a8b0ad1c8cd787e4e98449b71a2ac2d678", "submitted_at": "2026-08-25T03:16:45Z", "updated_at": "2026-08-25T03:16:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24017v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.24017", "identifiers": {"arxiv": "2608.24017"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24017"}], "name": "WebMCP-Phalanx: Enforcing and Characterizing Trust Boundaries for Browser-Integrated LLM Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-25", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Proposes WebMCP-Phalanx, a browser-integrated runtime that binds page-exposed tools to their registering principals and separates content inspection from privileged invocation. It studies lifecycle and prompt-injection attacks, including a timing bypass that motivates validating all visible metadata before tool use.", "topics": ["safety-security", "tool-integration"], "updated_at": "2026-08-25", "url": "https://arxiv.org/abs/2608.24017"}
{"aliases": [], "authors": ["Guo Gan", "Yilun Zhao", "Cong Chen", "Jinbiao Wei", "Tingyu Song", "Zheyuan Yang", "Lin Fu", "Hong Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51901009de5898c52a1b9d4958b698808a8f8b850d19afa32dd7a3f32d42a851", "submitted_at": "2026-08-25T05:57:25Z", "updated_at": "2026-08-25T05:57:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24099v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.24099", "identifiers": {"arxiv": "2608.24099"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24099"}], "name": "Are Android GUI Agents Robust Against Runtime Anomalies? AnTrap: Evaluating Agents in Dynamic Adversarial Environments", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AnTrap, which injects task-solvable runtime anomalies into Android GUI-agent trajectories using a layered failure taxonomy. It compares model robustness and adversarial reinforcement learning to distinguish recoverable interface traps from persistent contextual failures.", "topics": ["evaluation", "robustness", "reinforcement-learning"], "updated_at": "2026-08-25", "url": "https://arxiv.org/abs/2608.24099"}
{"aliases": [], "authors": ["Tao Xiong", "Xavier Hu", "Wenkai Wang", "Qinzhuo Wu", "Changqiao Wu", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76cd09f70bfec0d2a49cddae8adcf32ca37fb39bcade7c1d36d5f3dc3fa6bfa7", "submitted_at": "2026-08-25T07:40:12Z", "updated_at": "2026-08-25T07:40:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24174v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.24174", "identifiers": {"arxiv": "2608.24174"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24174"}], "name": "Task-Adaptive Rubrics for GUI Reward Modeling", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AdaptRubric, which retrieves reusable criteria for a GUI task family and then generates instruction-specific checks for concrete values and constraints. It evaluates these adaptive rubrics in offline reward assessment and online agent optimization.", "topics": ["reward-models", "evaluation", "reinforcement-learning"], "updated_at": "2026-08-25", "url": "https://arxiv.org/abs/2608.24174"}
{"aliases": [], "authors": ["Wenxuan Shen", "Dongna Jin", "Dongping Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f154c8f8d49f75fb249ebf1e2928a7db4d0568a3f9cc195284576f365438fbdf", "submitted_at": "2026-08-25T15:13:32Z", "updated_at": "2026-08-25T15:13:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24680v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.24680", "identifiers": {"arxiv": "2608.24680"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.24680"}], "name": "Game2World Engine: Unlocking In-the-Wild Gameplay Videos for World Model Training", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-25", "url": "https://arxiv.org/abs/2608.24680"}
{"aliases": [], "authors": ["Fei Tang", "Huawen Shen", "Zhiqiong Lu", "Zhengxi Lu", "Pengyuan Lyu", "Chengquan Zhang", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d2180e9b1b0c3ae20e32b0b597dc209e363551aaa5031c98504ef08314979a6", "submitted_at": "2026-08-25T17:35:42Z", "updated_at": "2026-08-25T17:35:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24848v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.24848", "identifiers": {"arxiv": "2608.24848"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24848"}], "name": "BrowserForge: Scaling Web Episode via Parallel Browser Sandboxes", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces BrowserForge, which combines broad website sourcing, parallel browser sandboxes, and task-proposer/solver interaction to collect verified web trajectories. The pipeline uses page structure during synthesis but trains a policy that acts from screenshots.", "topics": ["data-generation", "sandbox", "web-navigation"], "updated_at": "2026-08-25", "url": "https://arxiv.org/abs/2608.24848"}
{"aliases": [], "authors": ["Lily Bundgaard Wanscher", "Markus Heidemann Lorensen", "Mohammed Ammad Shafiq", "Mahyar Tourchi Moghaddam", "Mina Alipour"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "046d93a43dd5a011809a780c7d998f98ee0abb3e931d733b8ad7807b86050ec7", "submitted_at": "2026-07-26T17:21:24Z", "updated_at": "2026-07-26T17:21:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24913v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "id": "arxiv:2608.24913", "identifiers": {"arxiv": "2608.24913"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24913"}], "name": "From Blind Edits to Verified Repair: Building Trustworthy User-Side LLM Agents for Web Accessibility", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-07-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds a local browser-extension agent that generates and verifies additive CSS repairs for WCAG web accessibility violations.", "topics": ["ui-design", "web-navigation", "safety-security"], "updated_at": "2026-07-26", "url": "https://arxiv.org/abs/2608.24913"}
{"aliases": [], "authors": ["Min Ye", "Andrii Maksymov", "Nicolas Delfosse"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "179656ecbcaed01bd83bf863f2483dd2c30bedcb1f90ded8363f6f077e19eca3", "submitted_at": "2026-08-25T18:15:58Z", "updated_at": "2026-09-03T11:57:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.25027v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.25027", "identifiers": {"arxiv": "2608.25027"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.25027"}], "name": "Real-time decoder for a MegaQuOp quantum computer using a single CPU", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-03", "url": "https://arxiv.org/abs/2608.25027"}
{"aliases": [], "authors": ["Weiming Li", "Helen Paik", "Yulei Sui"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af6b60c6a39a990e950cf4b98f905f2d3aa5ae62d4055630c25a650508ec8bf8", "submitted_at": "2026-08-26T13:20:27Z", "updated_at": "2026-09-25T23:54:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.25777v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.25777", "identifiers": {"arxiv": "2608.25777"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.25777"}], "name": "LocalLSTC: A Long Short-Term Control Architecture for Locally Deployed GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces LocalLSTC, a training-free control architecture that maintains persistent task state and uses it to form short execution commitments. Outcomes feed back into that state for progress checks, recovery, and termination in locally deployed desktop agents.", "topics": ["state-tracking", "planning", "training-free"], "updated_at": "2026-09-25", "url": "https://arxiv.org/abs/2608.25777"}
{"aliases": [], "authors": ["Yuncheng Guo", "Zhanqiu Zhang", "Yiwen Guo", "Weijia Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "58f3e6aa42b8233eb5165d111ee12e40ef78963df7acb56cdd8608bee429f372", "submitted_at": "2026-08-25T17:21:15Z", "updated_at": "2026-09-10T16:56:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.26200v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.26200", "identifiers": {"arxiv": "2608.26200"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.26200"}], "name": "GameWAM: A World Action Model for Video Games", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-10", "url": "https://arxiv.org/abs/2608.26200"}
{"aliases": [], "authors": ["Joy Chen", "Alejandro Castillejo Munoz", "Pierluca D'Oro", "Yuxuan Sun", "Chloe Evans", "Joseph Tighe"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50c26a7ab741ecb3cb855305ee4f7d5b5729296248d650c469600aeb73101781", "submitted_at": "2026-08-25T23:21:57Z", "updated_at": "2026-09-15T04:43:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.26204v2"}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.26204", "identifiers": {"arxiv": "2608.26204"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.26204"}], "name": "ADeptS-Bench: Measuring the Trustworthiness of Computer Use Agents Across Devices", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates interface-embedded threats and whether computer-use agents seek clarification when mobile or desktop instructions are ambiguous.", "topics": ["safety-security", "evaluation", "human-in-the-loop"], "updated_at": "2026-09-15", "url": "https://arxiv.org/abs/2608.26204"}
{"aliases": [], "authors": ["Youtian Lin", "Yikang Yang", "Zhanpeng Hu", "Mengqi Zhou", "Feihu Zhang", "Xun Cao", "Jiaheng Liu", "Yao Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d019f053d03cafbfac691bce4fb167bdca1c6b11f9fefafedcff0e927eff2084", "submitted_at": "2026-08-26T17:54:39Z", "updated_at": "2026-08-26T17:54:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.26238v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.26238", "identifiers": {"arxiv": "2608.26238"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.26238"}], "name": "Procedura: Agentic 3D Modeling with Procedural Control", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-26", "url": "https://arxiv.org/abs/2608.26238"}
{"aliases": [], "authors": ["Zechun Niu", "Yukun Zhao", "Jiaxin Zhang", "Xu Shen", "Jinhua Si", "Han Tian", "Can Xu", "Yunfan Song", "Jiaxin Mao", "Yansong Gao", "Yuchen Li", "Jianmin Wu", "Lingyong Yan", "Shuaiqiang Wang", "Dawei Yin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ce34cc72ec72bd795e0efd83169e0da8151ae57989dd9beb4d290f7a8e65725", "submitted_at": "2026-08-27T02:36:23Z", "updated_at": "2026-08-27T02:36:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.26546v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.26546", "identifiers": {"arxiv": "2608.26546"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.26546"}], "name": "DuMateBench: Evaluating Autonomous Agents in Complex Real-World Workflows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-27", "url": "https://arxiv.org/abs/2608.26546"}
{"aliases": [], "authors": ["Rui Xie", "Lu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1626d5190f145c668c9b7f363b7f84904c12daae485e63b858e440ded5ea6df", "submitted_at": "2026-08-27T11:38:19Z", "updated_at": "2026-08-27T11:38:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.26991v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.26991", "identifiers": {"arxiv": "2608.26991"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.26991"}], "name": "ASIL: Replacing Screenshot-and-Click with Structured State and Semantic Actions", "notes": "Reviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.", "published_at": "2026-08-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "ASIL exposes application state as structured JSON and provides code-executable semantic actions. Its application benchmark compares this interface with screenshot control and application-native APIs, and studies supervised and reinforcement learning on the structured interface.", "topics": ["automation", "tool-integration", "evaluation", "reinforcement-learning", "imitation-learning"], "updated_at": "2026-08-27", "url": "https://arxiv.org/abs/2608.26991"}
{"aliases": [], "authors": ["Xingshan Zeng", "Zishan Xu", "Boju Zhang", "Yuzhou Wu", "Lingzhi Wang", "Jianghao Lin", "Liangyou Li", "Yasheng Wang", "Lifeng Shang", "Xin Jiang", "Weinan Zhang", "Yong Yu", "Qun Liu", "Weiwen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e2abdcb6a960364845631e0859fcc28a883251c0917c6ad5d3dd68be8bafee3a", "submitted_at": "2026-08-27T15:43:50Z", "updated_at": "2026-08-27T15:43:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27260v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.27260", "identifiers": {"arxiv": "2608.27260"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.27260"}], "name": "What Makes Good Agentic Data? An ACE Lens on Data Generation for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-27", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-27", "url": "https://arxiv.org/abs/2608.27260"}
{"aliases": [], "authors": ["Yiqi Zhu", "Feiyu Gao", "Jiaxing Fan", "Jiahui Zeng", "Minggang Wu", "Chenliang Li", "Haiyang Xu", "Peng Li", "Ming Yan", "Yang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a5ee0ab2cc4e3ae641c4ba27d64b7399e2f75d4b5533b2cdc42e3ace6a5dbdf", "submitted_at": "2026-08-21T12:17:14Z", "updated_at": "2026-08-21T12:17:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27477v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.27477", "identifiers": {"arxiv": "2608.27477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.27477"}], "name": "Benchmarking General Mobile Assistants in Challenging Real-World Scenarios", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-21", "url": "https://arxiv.org/abs/2608.27477"}
{"aliases": [], "authors": ["Yu Han", "Tianwen Qian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "576702fae164cbcb4adc86a81115487aae7e1c40cf07a1ecf3ef7c4c7e822c3b", "submitted_at": "2026-08-27T07:09:28Z", "updated_at": "2026-08-27T07:09:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27508v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.27508", "identifiers": {"arxiv": "2608.27508"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.27508"}], "name": "WM-R1: Training GUI Agents to Reason and leverage World Models with Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses learned world models for mobile-agent training rollouts and for reasoning about the consequences of candidate GUI actions.", "topics": ["world-models", "reinforcement-learning"], "updated_at": "2026-08-27", "url": "https://arxiv.org/abs/2608.27508"}
{"aliases": [], "authors": ["Divake Kumar", "Sina Tayebati", "Devashri Naik", "Amanda Sofie Rios", "Nilesh Ahuja", "Omesh Tickoo", "Ranganath Krishnan", "Amit Ranjan Trivedi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14864aae8f977d474ff3e7dd3e38951374862abfb8077395f2dc5a9e1cc5a489", "submitted_at": "2026-08-28T00:50:47Z", "updated_at": "2026-08-28T00:50:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27808v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.27808", "identifiers": {"arxiv": "2608.27808"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.27808"}], "name": "CURA: Certified Runtime Alarms for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CURA, a telemetry-based runtime monitor that turns computer-use trajectories into sequential failure alarms and can trigger mid-task oversight. Its certificate controls false alarms, while the study separately reports detection performance and limits of retrospective monitoring.", "topics": ["observability", "safety-security", "evaluation"], "updated_at": "2026-08-28", "url": "https://arxiv.org/abs/2608.27808"}
{"aliases": [], "authors": ["Jiahe Ying", "Wendong Bu", "Kaihang Pan", "Bingchen Miao", "Siyu Chen", "Wen Wang", "Xueming Jiang", "Juncheng Li", "Siliang Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "352699b04f3a9406d7cab95ce9b5fb000e3e98a28996777e75b7399a6290357a", "submitted_at": "2026-08-28T03:15:11Z", "updated_at": "2026-08-28T03:15:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27866v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.27866", "identifiers": {"arxiv": "2608.27866"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.27866"}], "name": "Iron: Intent-Aligned and Retrospective Dual Learning Framework for Enhancing Generalist Virtual Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Iron, a GUI-agent training framework that aligns low-level actions with high-level intent using cycle-consistency rewards. It also repurposes failed trajectories through hindsight reconstruction and evaluates transfer to other digital tasks.", "topics": ["reinforcement-learning", "data-generation", "grounding"], "updated_at": "2026-08-28", "url": "https://arxiv.org/abs/2608.27866"}
{"aliases": [], "authors": ["Dong-Yeop Na", "Somyeong Lee", "Chung Hyun Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b72b85257a47ae99244ea761428e7ea5fd629fe7b8f659e4a498a808963cc15", "submitted_at": "2026-08-28T05:26:54Z", "updated_at": "2026-08-28T05:26:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27936v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.27936", "identifiers": {"arxiv": "2608.27936"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.27936"}], "name": "Accelerated S-NFC for Million-Chaff RCS Computation Using Low-Rank Compression of Concatenated Block Rows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-28", "url": "https://arxiv.org/abs/2608.27936"}
{"aliases": [], "authors": ["Junxuan Li", "Zijun Liu", "Ziyi Huang", "Peng Li", "Yuzhou Liu", "Ming Yan", "Yang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bbf657b2228bad25e668e3c9d7ec20c0ecab403d34b7e0b5608d350c46b8e50f", "submitted_at": "2026-08-29T14:45:09Z", "updated_at": "2026-08-29T14:45:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.29305v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.29305", "identifiers": {"arxiv": "2608.29305"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.29305"}], "name": "Learning Simple Test-Time Environments for LLM Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Test-Time Environment Decomposition, which uses trial interaction to break complex web observations into simpler modules. A label-free adaptation procedure learns from these sub-environments and studies whether their gains compose in full web tasks.", "topics": ["planning", "robustness", "web-navigation"], "updated_at": "2026-08-29", "url": "https://arxiv.org/abs/2608.29305"}
{"aliases": [], "authors": ["Sanket Badhe", "Deep Shah", "Priyanka Tiwari", "Nehal Kathrotia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "13b6833243a543eb156ce7f4d943f8a3c608b3b38b7eaa77605ca7484cdedc43", "submitted_at": "2026-08-30T06:36:42Z", "updated_at": "2026-09-28T17:25:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.29596v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T13:00:09.204016+00:00", "role": "primary", "sha256": "d112d0a83472b87de538d0290efdd2d0dd0d21e878acca5f383da5baa880fe88", "supports": ["scope", "environments", "notes"], "url": "https://arxiv.org/html/2608.29596v2"}], "id": "arxiv:2608.29596", "identifiers": {"arxiv": "2608.29596"}, "kind": "survey", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.29596"}], "name": "A Systematic Survey of Agentic Skills: Architecture, Lifecycle, and Security", "notes": "Adjacent broad survey of procedural agent skills. Full-text Section 4.4.2 explicitly covers desktop, mobile, and web interface skills; surveyed systems and their reported results are not independently reproduced.", "published_at": "2026-08-30", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Surveys modular agentic skill architectures, skill lifecycles, and security risks in autonomous tool- and interface-using agents.", "topics": ["survey", "workflow", "memory", "safety-security"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2608.29596"}
{"aliases": [], "authors": ["Zhendong Li", "Lei Sun", "Letian Shi", "Deheng Zhang", "Ruibo Ming", "Mengshun Hu", "Dannong Xu", "Jian Wang", "Danda Paudel", "Luc Van Gool", "Jinjin Gu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6050ede322427b94aad748c324cae69b4c05f38561e8b191dde4d099de5face8", "submitted_at": "2026-08-30T14:28:50Z", "updated_at": "2026-08-30T14:28:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.29814v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.29814", "identifiers": {"arxiv": "2608.29814"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.29814"}], "name": "FRAMEWORKERS: A Dynamic Multi-Agent Framework for AI-Generated Video Production", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-30", "url": "https://arxiv.org/abs/2608.29814"}
{"aliases": [], "authors": ["Mingshuo Wang", "Hanqing Guo", "Huining Li", "Yuliang Fu", "Jing Xu", "Chenhan Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b17f858c9d6992f6a8d8f21cc79d468b0f1c81fc44084045791d9c3acacd6d0", "submitted_at": "2026-08-30T20:55:21Z", "updated_at": "2026-08-30T20:55:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.30038v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2608.30038", "identifiers": {"arxiv": "2608.30038"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.30038"}], "name": "ActReal: System-Level Mobile Agents Challenge Mobile Automation Detection", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ActReal, which translates semantic mobile-agent actions into coordinated touch and sensor signals to study automation-detection weaknesses. It evaluates detectors that observe touch events alone or together with inertial measurements under a privileged executor threat model.", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2026-08-30", "url": "https://arxiv.org/abs/2608.30038"}
{"aliases": [], "authors": ["Chen Xiong", "Zhiyuan He", "Pin-Yu Chen", "Stjepan Picek", "Tsung-Yi Ho"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87c3f58eebe2db437b1166ff8e83b336db4cde37684a7cd86f61c4b34fca7237", "submitted_at": "2026-08-31T03:39:43Z", "updated_at": "2026-10-03T10:41:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.30207v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2608.30207", "identifiers": {"arxiv": "2608.30207"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.30207"}], "name": "SIR: Self-improving Red-teaming for Compute Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SIR, which iteratively refines a reusable library of prompt-injection strategies from failed computer-use attack trajectories. Its evaluation uses deterministic system-state checks and counts attacks only when the adversarial objective and legitimate task both complete.", "topics": ["safety-security", "evaluation", "data-generation"], "updated_at": "2026-10-03", "url": "https://arxiv.org/abs/2608.30207"}
{"aliases": [], "authors": ["Xuehai Wang", "Haowei Qin", "Tongxin Liu", "Junkai Li", "Buqiang Xu", "Jintian Zhang", "Yijun Chen", "Zirui Xue", "Shumin Deng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4b21d6b2506ba47d6c4915f0643bab5f7527ca3cbe342eecd96c6e08440b33ea", "submitted_at": "2026-08-31T16:48:51Z", "updated_at": "2026-08-31T16:48:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.31076v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2608.31076", "identifiers": {"arxiv": "2608.31076"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2608.31076"}], "name": "Learning to Evaluate Before Improving: Automatic Rubric Induction for Automatic Research Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2608.31076"}
{"aliases": [], "authors": ["Venus Team", "Zhuohan Cai", "Haoxing Chen", "Jiaxuan Chen", "Weizhi Chen", "Changlong Gao", "Zhangxuan Gu", "Yuan Guo", "Yusong Hu", "Jianrong Jiang", "Jianguo Li", "Runze Li", "Jinzhen Lin", "Zhenyu Ma", "Changhua Meng", "Han Peng", "Xinyu Qiu", "Shuheng Shen", "Zhongyi Shui", "Weiqiang Wang", "Ming Wen", "Zhuoer Xu", "Hang Yan", "Kaiwen Yang", "Ruilin Yao", "Nanjun Yu", "Zhengwen Zeng", "Lianrui Zhang", "Yunzhu Zhang", "Zhe Zhao", "Beitong Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "34383fa6c6e5cca059c53fd0452d93eec9ac6a96271ddc14f81ddb96fdeaceec", "submitted_at": "2026-08-27T17:15:33Z", "updated_at": "2026-08-27T17:15:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.00028v1"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2609.00028", "identifiers": {"arxiv": "2609.00028"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.00028"}], "name": "UI-Venus-2 Technical Report", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a GUI foundation agent trained through expanded environments, function-grounded tasks, and trajectory verification with safety-aware execution mechanisms.", "topics": ["automation", "reinforcement-learning", "safety-security"], "updated_at": "2026-08-27", "url": "https://arxiv.org/abs/2609.00028"}
{"aliases": [], "authors": ["Lin Fu", "Zheyuan Yang", "Tianhui Zhang", "Jinbiao Wei", "Guo Gan", "Boxu Liu", "Yilun Zhao", "Yu Rong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dba8eb3f30a7e88ad973080d07fbf272f63400f4f8fa858d5f2657735ce093c0", "submitted_at": "2026-08-30T05:19:43Z", "updated_at": "2026-08-30T05:19:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.00048v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2609.00048", "identifiers": {"arxiv": "2609.00048"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.00048"}], "name": "GUI-CC: Benchmarking Contextual Consistency of GUI World Models as Agent Environments", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUI-CC, a benchmark for assessing GUI world models as multi-step agent environments. Separate reference-action and agent-loop tracks test whether generated screens preserve context, plausible transitions, and task progress throughout interaction.", "topics": ["world-models", "evaluation", "long-horizon"], "updated_at": "2026-08-30", "url": "https://arxiv.org/abs/2609.00048"}
{"aliases": [], "authors": ["Satwik Ram Kodandaram", "Monalika Padma Reddy", "Xiaojun Bi", "Jiawei Zhou", "I. V. Ramakrishnan", "Vikas Ashok"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "deb130d93bbbd6dfdc04c72d232a9dcfa1e92b0586831e5864070e7fe0de14dd", "submitted_at": "2026-09-01T00:44:19Z", "updated_at": "2026-09-01T00:44:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.00524v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.00524", "identifiers": {"arxiv": "2609.00524"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.00524"}], "name": "Are We There Yet? Assessing Computer-Use Agents for Blind Users' Accessible Interaction with Desktop Applications", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "A diary study deploys the screen-reader-accessible OLLA computer-use prototype with blind users across desktop applications. Recorded commands and interaction traces support model comparisons and analysis of grounding, planning, constraint tracking and termination failures.", "topics": ["evaluation", "human-in-the-loop", "grounding"], "updated_at": "2026-09-01", "url": "https://arxiv.org/abs/2609.00524"}
{"aliases": [], "authors": ["Sitong Pan", "Yipeng Shen", "Yilin Lu", "Caiwen Ding", "Lu Cheng", "Qianwen Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "967f1cdda69349da3280827c0b90301e4942c44398902ab6cfb868f1d3e94a84", "submitted_at": "2026-09-02T03:34:47Z", "updated_at": "2026-09-02T03:34:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02057v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.02057", "identifiers": {"arxiv": "2609.02057"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.02057"}], "name": "Monitoring Web Agents Without Internal Signals: Observable Trajectories and Key-Step Supervision", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Predicts web-agent failure risk from observable trajectory prefixes using cross-step behavior and black-box consistency queries. Key-step labels preserve valid early prefixes of failed runs, and evaluation tests early intervention and transfer across website categories.", "topics": ["observability", "evaluation", "robustness", "interaction-history"], "updated_at": "2026-09-02", "url": "https://arxiv.org/abs/2609.02057"}
{"aliases": [], "authors": ["Yixiong Xiao", "Lang An", "Hucheng Yang", "Pinxue Ma", "Yongquan Chen", "Jingjia Cao", "Yusai Zhao", "Ting Wang", "Ting Liu", "Siqi Bao", "Jingbo Zhou", "Hua Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52849171f507e126093411ba81dc580d460b84f831358b08bd57cb54ca41b665", "submitted_at": "2026-09-02T06:03:18Z", "updated_at": "2026-09-10T15:52:42Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.02149v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.02149", "identifiers": {"arxiv": "2609.02149"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.02149"}], "name": "OmegaUse-SOP: SOP Engineering for Professional Computer Use from Human Demonstrations", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "OmegaUse-SOP converts expert GUI demonstrations into configurable procedures and refines them with domain rules and human input. It records, abstracts, grounds and verifies live execution, with a case study of photovoltaic simulation workflows in PVsyst 7.2.", "topics": ["human-in-the-loop", "recording", "workflow", "automation"], "updated_at": "2026-09-10", "url": "https://arxiv.org/abs/2609.02149"}
{"aliases": [], "authors": ["Yimeng Liu", "Hua Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b462dddd45e6b04a3bf6fa0bc73576a23a5f9617f61aeb257a845c0f456f731d", "submitted_at": "2026-09-02T06:06:43Z", "updated_at": "2026-09-02T06:06:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02154v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2609.02154", "identifiers": {"arxiv": "2609.02154"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.02154"}], "name": "WiP: Characterizing and Defending Against Mobile-Agent-Driven MFA Automation", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies how mobile automation can combine authentication factors within one agent environment and evaluates motion sensing as a user-presence signal.", "topics": ["safety-security", "evaluation"], "updated_at": "2026-09-02", "url": "https://arxiv.org/abs/2609.02154"}
{"aliases": [], "authors": ["Ihor Stepanov", "Aleksandr Smechov", "Mykhailo Shtopko", "Dmytro Vodianytskyi", "Oleksandr Lukashov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5552726e5b7890ba914c132bc62e0e16a58acd78e8e537fc892b797e7041c726", "submitted_at": "2026-09-02T08:40:07Z", "updated_at": "2026-09-02T08:40:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02292v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.02292", "identifiers": {"arxiv": "2609.02292"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.02292"}], "name": "SCX Router: Streaming Zero-Shot Model Selection with a Decoder-KV Classifier and a Real-World Task Ontology", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-02", "url": "https://arxiv.org/abs/2609.02292"}
{"aliases": [], "authors": ["Bizhe Bai", "Jiakang Yuan", "Hongming Wu", "Xinyue Wang", "Jie Ren", "Siyao Chen", "Yuchen Ya", "Fan Bai", "Pai Peng", "Huafeng Qin", "Tao Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ec439fd2cfdf7319d77ef5ab58e50143bceb3f8205ccffa0d83555f10dece42a", "submitted_at": "2026-09-02T08:54:58Z", "updated_at": "2026-09-02T08:54:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02309v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.02309", "identifiers": {"arxiv": "2609.02309"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.02309"}], "name": "Efficient GUI Agents: A Systems Survey of Observation, Memory, Action, and Runtime Optimization", "notes": "", "published_at": "2026-09-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Surveys GUI-agent efficiency beyond task success: observation, context and memory, action, and planner or runtime cost.", "topics": ["survey"], "updated_at": "2026-09-02", "url": "https://arxiv.org/abs/2609.02309"}
{"aliases": [], "authors": ["Kelvin Li", "Dhruv Pendharkar", "Anish Pahilajani", "Chuyi Shang", "Leon Oks", "Leonid Karlinsky", "Rogerio Feris", "Trevor Darrell", "Roei Herzig"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8703474cb4f577b307731cf1a986f76f800470fde625b6ca4b3569d30e74f903", "submitted_at": "2026-09-02T17:59:40Z", "updated_at": "2026-09-02T17:59:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02885v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.02885", "identifiers": {"arxiv": "2609.02885"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.02885"}], "name": "Discriminative World Models for Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains discriminative web world models to distinguish the true next state from outcomes of alternative actions. Branching web trajectories support training, followed by tests of predicted-state matching, process-reward action ranking and end-to-end WebArena-Lite selection.", "topics": ["world-models", "reward-models", "planning", "data-generation", "evaluation"], "updated_at": "2026-09-02", "url": "https://arxiv.org/abs/2609.02885"}
{"aliases": [], "authors": ["Zhaoyuan Huang", "Tianjie Ju", "Pengzhou Cheng", "Zheng Wu", "Yansi Li", "Chuanbiao Song", "Jun Lan", "Huijia Zhu", "Weiqiang Wang", "Zhuosheng Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a59f7d54845de6efebf6365a188e61e41fdf31e9ebe894b1d01ce7b86d6be4ec", "submitted_at": "2026-09-03T06:48:41Z", "updated_at": "2026-09-03T06:48:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.03438v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.03438", "identifiers": {"arxiv": "2609.03438"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.03438"}], "name": "Do GUI Agents Know When Not to Act? Enabling Conflict-Aware Termination for Multimodal GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "CONFLICTGUI evaluates termination under instruction conflicts and mismatches between instructions and interface context. CONFLICTGUARD couples feasibility checks with conditional action modulation to reduce unnecessary execution while tracking normal-task behavior.", "topics": ["safety-security", "evaluation", "robustness", "planning"], "updated_at": "2026-09-03", "url": "https://arxiv.org/abs/2609.03438"}
{"aliases": [], "authors": ["Vilém Zouhar", "Niyati Bafna", "Mukund Choudhary", "Maike Züfle", "Sara Rajaee", "Pinzhen Chen", "Jannis Vamvas", "Sara Papi", "Ona de Gibert", "Bhavitvya Malik", "Eliya Habba", "Orfeas Menis Mastromichalakis", "Patrícia Schmidtová", "Michelle Wastl", "Sheriff Issaka", "Leshem Choshen", "Stella Biderman", "Antonis Anastasopoulos", "Jan Niehues", "Rico Sennrich", "Mrinmaya Sachan", "Ondřej Bojar", "Kenton Murray", "Jörg Tiedemann", "Alham Fikri Aji", "Philipp Koehn", "Christof Monz", "Alexandra Birch", "Sowmya Vajjala", "Chalamalasetti Kranti", "Cristina España-Bonet", "Nobin Sarwar", "David Kaczér", "Sourajit Saha", "Jonathan Tonglet", "Shunta Asano", "Malik Marmonier", "Daban Q. Jaff", "Vaisakhi Mishra", "Hend Al-Khalifa", "Gabriele Sarti", "Nils Rehlinger", "Juan Daniel Cuervo Villa", "Dominik Macháček", "Saugata Purkayastha", "Jagannathan Ramanujam", "Shubhashis Roy Dipta", "Aviral Nigam", "Shuaib Shuaib Yusuf", "Heejin Do", "Jonathan Yahav", "Johannes-Rudolf David", "Maria Carmen Staiano", "Zuzana Nadova", "Fred Philippy", "Ron Keinan", "Maria Lymperaiou", "Silvia Casola", "Fabian Retkowski", "Andrés Jerez", "Hanna Yukhy- menko", "Sangwon Ryu", "Avantica Vempati", "Sukannya Purkayastha", "Adrian Cosma", "Erivan Inan", "Vitalii Babenko", "Wafa Aissa", "Valentin Scourneau", "Fatima Haouari", "Venkata Prasanth Kumar Gummadi", "Mehdi Jafarzadeh", "Manon Reusens", "Kaiser Sun", "Lukas Edman", "Shaomu Tan", "Giuseppe Gallipoli", "Pawan Sasanka Ammanamanchi", "Manar Ali", "Mohammad Sadegh Gholizadeh", "Dipankar Srirag", "Marek Šuppa", "Javier García Gilabert", "Ruta Binkyte", "Ana-Maria Bucur", "Sabry E. Farrag", "Youssef Saber", "Yihong Liu", "Theresia Veronika Rampisela", "Christian Hoang", "Nicoleta Cojocaru", "Jan Kocoń", "Jean Maillard", "Xiaochuang Yuan", "Sina Ahmadi", "Daryna Dementieva", "Philipp Mondorf", "Kaustubh Dhole", "Valmik Nahata", "Roman Wixinger", "Amir Hossein Yari", "Shenbin Qian", "Manuel Tuor", "Fida Mohammad Thoker", "Sergey Troshin", "Lance Calvin Lim Gamboa", "Amir Arsalan Rezapour", "Kätriin Kukk", "Koel Dutta Chowdhury", "Shaswati Saha", "Seth Aycock", "Bo Chen", "Linh Vu", "Vatsal Venkatkrishna", "Shayan Bali", "Arafat Ahsan", "Luan Thanh Nguyen", "Hassan Soliman", "Ngoc Quynh Tram Do", "Azmine Toushik Wasi", "R. Damanhuri", "Marius Huber", "Kazuki Egashira", "Jimson Paulo Layacan", "David Africa", "Vladislav Poritski", "Mike Zhang", "Deep Shah", "Abdulaziz Nura Kani", "Luis Frentzen Salim", "Paul Gavrikov", "Bello Umar Bello", "Ayush Sunil Munot", "Anumit Garg", "Yolanda Xavier", "Qiaoyuan Zheng", "Kawsar Ahmed", "Debanshu Das", "Zimu Wang", "Gengyu Rao", "Kamile Dementaviciute", "Farhan Farsi", "L D M S Sai Teja", "Dawei Zhu", "Yi Fan", "Wei Liu", "Marco Gaido", "Elias Herranen", "Sankalan Pal Chowdhury", "Karen Sanchez", "Guy Kaplan", "Farzad Shami", "Ashok Urlana", "Amir Hossein Kargaran", "Sofie Goethals", "Priyaranjan Pattnayak", "Oksana Volchek", "Marii Ojastu", "Hongbin Na", "Emilian Radoi", "Chenyi Zhao", "Carlos Hinojosa", "Andrei Niculae", "Andrea Gregor de Varda", "Zaid Alyafeai", "Tomasz Limisiewicz", "Reem Alzahrani", "Pouya Sadeghi", "Nehal Kathrotia", "Mateusz Lango", "Enzo Doyen", "Alex Flückiger", "Ulysses Sekai Tully Carr", "Samuel Simko", "Ritwik Tiwari", "Rishit Dagli", "Isaac R Caswell", "Bowen Yi", "Aicha Chorana", "Selja Keränen", "Sadiksha Chitrakar", "Muhammad Ravi Shulthan Habibi", "Joy Olusanya", "Bishal Shrestha", "Zhengxiang Wang", "Vivek Harsha Lakkamaneni", "Sophia Conrad", "Panayiotis Panayiotou", "Nazia Tasnim", "Marta Punsola Munárriz", "Marko Culjak", "Luis Lara", "Jenny Chim", "Jannatul Nayem", "Fidel Rodríguez Velásquez", "Eran Yahav", "Blanka Kövér", "Beatrice Savoldi", "Anmol Goel", "Aishik Mandal", "Tosin Adewumi", "Raoyuan Zhao", "Mykola Haltiuk", "Antonia Karamolegkou", "Yuxing Lu", "Thura Aung", "Naser Almousa", "Tommaso Cerruti", "Raia Abu Ahmad", "Beni Egressy", "Alireza Pakniat", "Stéphane J. P. S. Thunus", "Rachel Bawden", "Lena Libon", "Samridhya Biswas", "Prakhar Gupta", "Nusrat Jahan Lia", "Nguyen Tai", "Natchapon Jongwiriyanurak", "Minh Ngoc Do", "Ivan Barać", "Dzmitry Kuzmin", "Badal Nyalang", "Antoine Taroni", "Andy Catruna", "Rushikesh Zawar", "Roland Aydin", "Pavel Stepachev", "Ilai Yaron Levy", "Andreas Simons", "Rayyan Merchant", "Ziyi Yang", "Samuel Frontull", "Kenneth Enevoldsen", "Harris Abdul Majid", "Tim Graf", "Tatiana Bielakova", "Sharifa Djurabaeva", "Shaoxiong Ji", "Jirui Qi", "Ayla Rigouts Terryn", "Yurii Paniv", "Xiyan Fu", "Sunisth Kumar", "Shree Harsha Bokkahalli Satish", "Papa Abdou Karim Karou Diallo", "Mengyu Ye", "Maximilian Vieweg", "Matija Akrap", "Kristýna Onderková", "Joseph Attieh", "Ivan Yuri De Leon", "Ibrahim Baroud", "Esrael Teferi Tensay", "Elisabeth Fittschen", "David Dukić", "Benoît Sagot", "Jingwei Ni", "Yu Fan", "Juri Opitz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6b8995fda8aaf8a5d13d3fffe19593c841fb8e1e6f56ab197f51aa0e6fed9f1", "submitted_at": "2026-09-03T17:54:45Z", "updated_at": "2026-09-29T06:41:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.04173v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.04173", "identifiers": {"arxiv": "2609.04173"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.04173"}], "name": "Last Translation Benchmark", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.04173"}
{"aliases": [], "authors": ["Lin Shi", "Haowei Lin", "Zixuan Zhu", "Xiaoyue Zhou", "Xiang Li", "Xiangning Lin", "Yaxuan Deng", "Han Xu", "Yuangang Li", "Shanda Li", "Zizhao Chen", "Hanwen Xing", "Harsh Raj", "Bo Chen", "Quan Shi", "Steven Dillmann", "Yipeng Gao", "Puneesh Khanna", "Ruofan Lu", "Chao Beyond Zhou", "Michael Yang", "Robert Zhang", "Siyuan Chai", "Jiayu Chang", "Yizhao Chen", "Xiaokun Chen", "Yiwei Dai", "Wenting Yang", "Hange Liu", "Minghao Liu", "Zihan Wang", "Adnan El Assadi", "Benedikt Stroebl", "E. Kelly Buchanan", "Han Meng", "Junwei He", "Longxuan Yu", "Radin Shayanfar", "Yukyung Lee", "Zhikang Dong", "Allen G Hart", "Anjiang Wei", "Anurag Kashyap", "Arpandeep Khatua", "Audrey Jixin Zheng", "Chengrui Ma", "David Heineman", "Dubing Chen", "Hai-Anh Trinh", "Haishuo Fang", "Hefan Zhang", "Hui Shen", "Issa Sugiura", "Jiankai Sun", "Jiechao Gao", "Junhong Lin", "Junnan Li", "Kai Yang", "Lei Hsiung", "Maoyu Wang", "Mengze Tang", "Nabil Omi", "Negin Raoof", "Nicholas Edwards", "Octavia Guo", "Orfeas Menis Mastromichalakis", "Pengliang Ji", "Przemysław Hejman", "Qi Qi", "Qunshu Lin", "Richard Zhuang", "Rui Yang", "Ruichen Zheng", "Ryan Marten", "Shaghayegh Fazliani", "Shizheng Hou", "Sicong Jiang", "Sijie Li", "Boqin Yuan", "Michael Glass", "Song Bian", "Terry Yue Zhuo", "Tianqing Wu", "Tom Tang", "Wanjia Zhao", "Weihao Xuan", "Wenhua Liang", "Xian Liu", "Xin Lan", "Xuan Zhang", "Xuandong Zhao", "Yanchuan Tang", "Yifan Jiang", "Yijiang Li", "Yitong Guan", "Yizhi Li", "Yonghui Liu", "Yuheng Tang", "Yujun", "Mao", "Yunfei Zhao", "Yuxin Wang", "Yuxuan Tang", "Zhenheng Tang", "Zhifei Li", "Ziruo Wang", "Ziyu She", "Kaiyuan Liu", "Iheb Chaabane", "Yuxin Tang", "Xiangyi Li", "Satya Sai Srinath Namburi GNVV", "Xinyue Zheng", "Andy Konwinski", "Boxuan Li", "Leon Liangyu Chen", "Alex Dimakis", "Nicholas Carlini", "Soroush Vosoughi", "Sanmi Koyejo", "Di He", "Etash Guha", "Benjamin Feuer", "Mike Merrill", "Ludwig Schmidt", "Alex Shaw"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8d8fd6b0f47df7384e0ca1c604c0f667b15bd3b683147df035dd8f36dde6337", "submitted_at": "2026-09-03T16:26:20Z", "updated_at": "2026-09-10T02:39:50Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.04298v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.04298", "identifiers": {"arxiv": "2609.04298"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.04298"}], "name": "Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agentic Evaluation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-10", "url": "https://arxiv.org/abs/2609.04298"}
{"aliases": [], "authors": ["Weide Zhan", "Qumu Shaqu", "Yuanqing Liu", "Peng Zhang", "Jiahao Liu", "Kam Him Lam", "Ning Gu", "Zhan Hu", "Tun Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61587798aefd8e9592ddaef756a13e02f0e4a987e4068ef582d0aaf66e8eb680", "submitted_at": "2026-09-04T08:09:47Z", "updated_at": "2026-09-04T08:09:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.04850v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2609.04850", "identifiers": {"arxiv": "2609.04850"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.04850"}], "name": "ElderBench: Benchmarking Autonomous Mobile Agents for Older Adults", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "ElderBench evaluates mobile GUI agents on smartphone requests elicited from older adults, including indirect and underspecified language. Online and offline tests, instruction normalization and linguistic analyses examine how these request patterns contribute to failures.", "topics": ["evaluation", "human-in-the-loop", "mobile-navigation", "robustness"], "updated_at": "2026-09-04", "url": "https://arxiv.org/abs/2609.04850"}
{"aliases": [], "authors": ["Longtao Hu", "Xiao Liang", "Linchao Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "78db974038e0dde9a1347132382573fa9e2aafb1b0e1a574da3d769574329be9", "submitted_at": "2026-09-04T08:29:45Z", "updated_at": "2026-09-04T08:29:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.04869v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.04869", "identifiers": {"arxiv": "2609.04869"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.04869"}], "name": "From Interaction Traces to Persistent Skills: Online Evolution for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "An online skill-evolution framework turns GUI interaction traces and evaluator feedback into a versioned library of reusable procedures without updating model parameters. Controlled OSWorld comparisons examine both the benefits of shared procedural memory and its domain-dependent instability under repeated revision.", "topics": ["memory", "workflow", "evaluation", "training-free"], "updated_at": "2026-09-04", "url": "https://arxiv.org/abs/2609.04869"}
{"aliases": [], "authors": ["Haoting Shi", "Wenhao Wang", "Weicheng Fang", "Yaozhong Liang", "Tian Jin", "Pengxiang Zhao", "Guangyi Liu", "Siheng Chen", "Yanfeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "69eab19b6e45e1b2ddf9f29f6f0937d51e06a5bb7dea34dadb761361c82f08d4", "submitted_at": "2026-09-04T17:26:18Z", "updated_at": "2026-09-04T17:26:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05374v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2609.05374", "identifiers": {"arxiv": "2609.05374"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.05374"}], "name": "CUA-Universe: A Scalable and Dynamic Environment for Hybrid GUI+CLI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Turns desktop applications into reproducible GUI-and-CLI environments and generates verified hybrid trajectories for agent post-training.", "topics": ["tool-integration", "data-generation", "sandbox"], "updated_at": "2026-09-04", "url": "https://arxiv.org/abs/2609.05374"}
{"aliases": [], "authors": ["Bowei He", "Xiaokun Zhang", "Meng Ding", "Xue Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "edbe52f304634dfa06e0f232090a4fe998c7ea111418f57942976d348c4e0997", "submitted_at": "2026-08-31T08:16:56Z", "updated_at": "2026-08-31T08:16:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05511v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.05511", "identifiers": {"arxiv": "2609.05511"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.05511"}], "name": "SCAFFOLD: Self-Improving Web Agents via Recursive Parametric Skill Abstraction", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SCAFFOLD, which extracts executable skills from web trajectories, composes them into a hierarchy, and compresses redundant skills through behavioral checks. It periodically distills skill-guided trajectories into the policy and studies repeated improvement across web benchmarks.", "topics": ["memory", "automation", "web-navigation"], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2609.05511"}
{"aliases": [], "authors": ["Jianwei Zhang", "Sihan Cao", "Pengcheng Zheng", "Ya Wen", "Pei Ke", "Kuien Liu", "Shen Gao", "Wei Dong", "Yang Yang", "Chaoning Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aaeb2a57b2bb92ed926c9b65fb40b6bbf272e0ad9bb7415716d24eab0aeb4402", "submitted_at": "2026-08-31T08:40:34Z", "updated_at": "2026-08-31T08:40:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05513v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.05513", "identifiers": {"arxiv": "2609.05513"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.05513"}], "name": "When and What to Teach: Budget-Aware Online Adaptation for Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-31", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes an online teaching framework that decides which web-agent episodes warrant a stronger teacher and which trajectory turns merit training. A solvability gate and score-guided trimming aim to preserve first-attempt performance while reducing teaching and training costs.", "topics": ["efficiency", "data-generation", "web-navigation"], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2609.05513"}
{"aliases": [], "authors": ["Suyoung Lee", "Myungsub Choi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0a3b1dc46d9a875e2cfb2f9ec40dba631c999a6088b6d39d755239945f59d74", "submitted_at": "2026-09-02T07:03:03Z", "updated_at": "2026-09-02T07:03:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05535v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2609.05535", "identifiers": {"arxiv": "2609.05535"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.05535"}], "name": "Beyond the Verdict: Evidence-Aligned Evaluation of Visual Prompt-Injection Guardrails", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-02", "url": "https://arxiv.org/abs/2609.05535"}
{"aliases": [], "authors": ["Jiangyun Zhang", "Kristen Surrao", "Torpong Nitayanont", "Yupei Zhang", "Roopali Singh", "Zhiyu Chen", "Julia Huang", "Zhou Tang", "Shayan Ali Akbar", "Omar Alonso", "Erwin Cornejo", "Yuan Li", "Yi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43c19c25e0ca5210547c1adc16ad260a5d749a94c9aa0d4dadfe094f0d9b0501", "submitted_at": "2026-09-04T23:53:21Z", "updated_at": "2026-09-04T23:53:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05776v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.05776", "identifiers": {"arxiv": "2609.05776"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.05776"}], "name": "DI-Bench: Systematically Generating In-Domain Data Intelligence Benchmarks for Enterprise Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-04", "url": "https://arxiv.org/abs/2609.05776"}
{"aliases": [], "authors": ["Yu Liu", "Zhilin Liu", "Zhiwei Yang", "Shaojie Zhang", "Zheyuan Deng", "Tingwei Huang", "Zhenbo Luo", "Lei Jiang", "Yanbing Liu", "Pei Fu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "657a9c8f028c4fe00faef2e2830ca2ae49746a539d47e0f985e053e0225197fd", "submitted_at": "2026-09-05T12:36:59Z", "updated_at": "2026-10-14T06:29:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.06059v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.06059", "identifiers": {"arxiv": "2609.06059"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.06059"}], "name": "OmniStepBench: Deployment-Aware Agent Evaluation across Modalities and Task Complexity", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2609.06059"}
{"aliases": [], "authors": ["Zirui Shang", "Xin Shu", "Yang Liu", "Zhi Gao", "Xinxiao Wu", "Lifeng Fan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0d6c92c1bbf777119fdce5e9c4c5ded744c3459f8ae0b10ccc0fc75de11b237", "submitted_at": "2026-09-06T10:50:55Z", "updated_at": "2026-09-06T10:50:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.06530v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.06530", "identifiers": {"arxiv": "2609.06530"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.06530"}], "name": "Selective Knowledge Control for Continual GUI Agent Learning over Application Streams", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces activation-conditioned gradient control for GUI agents learning successive applications. It protects inactive application-specific neurons, constrains updates to shared neurons, and accumulates a compact historical state to balance adaptation with retention.", "topics": ["memory", "robustness"], "updated_at": "2026-09-06", "url": "https://arxiv.org/abs/2609.06530"}
{"aliases": [], "authors": ["Juyong Lee", "Woogyeol Jin", "Kimin Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14790f9565861cdba0b24661f6bfe87d7572cbad9d76d9a0b2d7216e9bac7ea7", "submitted_at": "2026-09-06T19:22:53Z", "updated_at": "2026-09-06T19:22:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.06792v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.06792", "identifiers": {"arxiv": "2609.06792"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.06792"}], "name": "Improving Proficiency and Efficiency of Android GUI Agents via Self-Generating Tool Actions", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "DroidTool generates application-state tools through proposal, implementation, testing and repair. Relational tests connect the tools’ preconditions and effects, and the resulting tools are evaluated as additions to Android GUI agents.", "topics": ["tool-integration", "automation", "mobile-navigation", "efficiency"], "updated_at": "2026-09-06", "url": "https://arxiv.org/abs/2609.06792"}
{"aliases": [], "authors": ["Boyang Zhang", "Qingxin Xiao", "Lingwei Dang", "Qingyao Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "58d19ba99e44c9155eb806de5621d344dcd2294c21debe568a57ef388b4c325a", "submitted_at": "2026-09-07T14:09:11Z", "updated_at": "2026-09-30T07:31:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.07529v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.07529", "identifiers": {"arxiv": "2609.07529"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.07529"}], "name": "CoER: Defending against Adaptive Indirect Prompt Injection via Adversarial Co-Evolution and Refinement", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2609.07529"}
{"aliases": [], "authors": ["Jingpu Yang", "Fengxian Ji", "Jinri Guo", "Tianhao Li", "Qian Jiang", "Fan Zhang", "Min Peng", "Qianqian Xie", "Preslav Nakov", "Zhuohan Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08972b7b01a189a42be94363db08d89c6f97bdcbd9e10860168b122cef29e8fa", "submitted_at": "2026-09-07T15:13:23Z", "updated_at": "2026-09-07T15:13:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.07603v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.07603", "identifiers": {"arxiv": "2609.07603"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.07603"}], "name": "FinCUABuild: Can Agents Build Reliable Benchmarks for Dynamic Financial Computer Use?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-07", "url": "https://arxiv.org/abs/2609.07603"}
{"aliases": [], "authors": ["Jintian Feng", "Long Chen", "Xiao Yu", "Jiayi Dai", "Chenglong Liu", "Haoru Wang", "Zizhen Xue", "Yuxuan Shi", "Ziyang Wang", "Yichen Gong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4090a2c7e7730ea415564b6b418d85aad77a74a222cef35435b0aa93beafd7f", "submitted_at": "2026-09-07T16:25:13Z", "updated_at": "2026-09-07T16:25:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.07712v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.07712", "identifiers": {"arxiv": "2609.07712"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.07712"}], "name": "APPSim-Bench: Bridging Real-world Apps and Reproducible Evaluation for Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces controlled simulated mobile apps with deterministic backend data and outcome checks for reproducible evaluation of realistic interactions.", "topics": ["evaluation", "sandbox", "mobile-navigation"], "updated_at": "2026-09-07", "url": "https://arxiv.org/abs/2609.07712"}
{"aliases": [], "authors": ["Pujun Zheng", "Zixin Shang", "Shufan Jiang", "Wenhui Tian", "Dongsheng Zhu", "Zerun Ma", "Dingbo Yuan", "Qi Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "41d3f82a933c2c8ce26e5e2113b04cec8801a18d6a8227d56ea0fe3cf533adf0", "submitted_at": "2026-09-08T02:29:58Z", "updated_at": "2026-10-09T02:16:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.08149v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.08149", "identifiers": {"arxiv": "2609.08149"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.08149"}], "name": "SWE-Bench Pro Verified: A Reliable Benchmark for Software Engineering Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2609.08149"}
{"aliases": [], "authors": ["Zongjie Li", "Alan Z. W", "John Nicolas J", "Walter H. F", "Scott Donald L", "Gordon Y. P", "Deke X"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "42ab51b07d42fa93e591266fd559d51c132f9f663665358a4f30cb1051a1d8ad", "submitted_at": "2026-09-08T08:26:43Z", "updated_at": "2026-09-08T08:26:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.08418v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.08418", "identifiers": {"arxiv": "2609.08418"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.08418"}], "name": "Feyospace-v1: How the Cyber Mercury Seven Trained Frontier Cyber Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-08", "url": "https://arxiv.org/abs/2609.08418"}
{"aliases": [], "authors": ["Xingyuan Bu", "Chengru Song", "Hao Zhou", "Tao Zhou", "Dong Li", "Wei Li", "Shilong Li", "Hao Shi", "Yongxin Guo", "Donghao Zhou", "Qiangpeng Yang", "Shilei Wen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5855f657f36e2d89e9838a460ea6bd5379bab7355c3f79b10a30756c1108336", "submitted_at": "2026-08-31T07:35:35Z", "updated_at": "2026-08-31T07:35:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.09187v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.09187", "identifiers": {"arxiv": "2609.09187"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.09187"}], "name": "AgenticGen: Reward-Guided Agentic Video Generation for Advertising", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-31", "url": "https://arxiv.org/abs/2609.09187"}
{"aliases": [], "authors": ["Zhihao Liu", "Hongyu Sun", "Zhiyuan Fu", "Xiaonan Duan", "Jice Wang", "Shangru Zhao", "Weizhi Meng", "Wuxin Yang", "Yangfan Zhou", "Yuqing Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a3a7a81ffe43885d9c8362df4c8aff08000304fe60477f484a732caf35124b2", "submitted_at": "2026-09-06T03:49:45Z", "updated_at": "2026-09-06T03:49:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.09212v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.09212", "identifiers": {"arxiv": "2609.09212"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.09212"}], "name": "AgentHijack: Visual Patch Attacks on Multimodal Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "AgentHijack evaluates whether visual patch effects propagate from screenshots through model outputs and action parsing into real environment changes. Its controlled web-page experiments distinguish intermediate attack signals from verified end-to-end consequences.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-09-06", "url": "https://arxiv.org/abs/2609.09212"}
{"aliases": [], "authors": ["Xiaolei Li", "Jialun Cao", "Zhijian Hou", "Yuzhi Zhao", "Yepang Liu", "Shing-Chi Cheung"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3307d90ef09c16afaee160eacc9d2994d142be9b9a26ce9c4a57f61345e2e8ec", "submitted_at": "2026-09-09T11:04:53Z", "updated_at": "2026-09-09T11:04:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.10031v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "id": "arxiv:2609.10031", "identifiers": {"arxiv": "2609.10031"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.10031"}], "name": "GraphDroid: Asynchronous LLM-Based Mobile App GUI Testing via History-Aware Exploration and Hybrid Intent Fulfillment", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "GraphDroid combines clustered memory of visited GUI states with asynchronous test-intent generation. A hybrid executor reserves LLM calls for complex intents and uses heuristics for simpler ones, evaluated through Android app coverage, bug discovery and crash detection.", "topics": ["evaluation", "mobile-navigation", "memory", "interaction-history", "efficiency"], "updated_at": "2026-09-09", "url": "https://arxiv.org/abs/2609.10031"}
{"aliases": [], "authors": ["Yuhao Wang", "Mu Qiao", "Xindong Zhang", "Yunzhi Zhuge", "Lei Zhang", "Huchuan Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "20d3bc3137e163f6b467b8c464101478b2375fbab8aebe9179e2e1113b34f503", "submitted_at": "2026-09-09T15:12:48Z", "updated_at": "2026-09-09T15:12:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.10297v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.10297", "identifiers": {"arxiv": "2609.10297"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.10297"}], "name": "TRACE: Trajectory-robust Admission with Evidence Ordering for Efficient GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "TRACE orders screenshot evidence for monotonic visual-token reduction using layout priors, instruction relevance and feature novelty. Spatial token reserves and incremental KV contraction aim to preserve useful evidence for future GUI targets without repeatedly encoding retired frames.", "topics": ["efficiency", "interaction-history", "screen-understanding", "training-free"], "updated_at": "2026-09-09", "url": "https://arxiv.org/abs/2609.10297"}
{"aliases": [], "authors": ["Zixiang Chen", "Yuheng Lu", "Zihao Cheng", "Zeming Liu", "Jizeng Bai", "Ziye Huang", "Zhiyin Lin", "Zihan Li", "Yuhang Guo", "Yunhong Wang", "Haifeng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5ec17516bce87a8e349983acac14415a6ceeb12c167fce57aa37d6157cd3cd83", "submitted_at": "2026-09-09T16:59:59Z", "updated_at": "2026-09-09T16:59:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.10451v1"}}, "environments": ["desktop", "android", "windows", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.10451", "identifiers": {"arxiv": "2609.10451"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.10451"}], "name": "JarvisGUI: Towards Cross-Device GUI Agents with Dynamic Task Composition", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "JarvisGUI composes typed input-output task transformations into dynamic workflows spanning Android, Windows and Ubuntu devices. Its virtual-environment benchmark tests transfer of intermediate results, shared state and cross-device dependencies.", "topics": ["evaluation", "cross-app", "planning", "long-horizon", "state-tracking"], "updated_at": "2026-09-09", "url": "https://arxiv.org/abs/2609.10451"}
{"aliases": [], "authors": ["Trevor Cappallo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ef89921b3d63b1e9fcc296a2018f2d96fadb6ef148a9d380e5ba06370836a92", "submitted_at": "2026-09-06T18:21:27Z", "updated_at": "2026-09-15T21:01:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.10592v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.10592", "identifiers": {"arxiv": "2609.10592"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.10592"}], "name": "Tag-system computation using only powers and principal logarithms", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-15", "url": "https://arxiv.org/abs/2609.10592"}
{"aliases": [], "authors": ["Sizhe Zhao", "Haozhe Xie", "Weiyu Zhao", "Chenchu Zhang", "Huan Wang", "Chenyang Wang", "Qinglin Liu", "Shengping Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6459d998a9621ec2b104c6f9f9546c19ca5a6a13e28165cf424d05dd8afe1598", "submitted_at": "2026-09-10T13:52:51Z", "updated_at": "2026-09-10T13:52:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.11561v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.11561", "identifiers": {"arxiv": "2609.11561"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.11561"}], "name": "Memory as Plans: World-Action Modeling with Memory-Grounded Planning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-10", "url": "https://arxiv.org/abs/2609.11561"}
{"aliases": [], "authors": ["Tong Ye", "Kunyang Han", "Guozhi Wang", "Longqiang Luo", "Zhifeng Ding", "Yongxiang Zhang", "Xiaolei Shen", "Yuxuan Zhang", "Zhuping Zhang", "Tao Xu", "Yue Pan", "Yucheng Zhao", "Yupei Hu", "Yuanjiang Ouyang", "Danfeng Shen", "Runqi Lin", "Hongda Cai", "Zhaoxiong Wang", "Mengjia Yan", "Yingjie Zhong", "Chen Zhou", "Zeyu Zhang", "Xuwen Zhu", "Penggang Shi", "Mingcheng Luo", "Ziyang Wu", "Min Jin", "Mingfu Shen", "Zairong Xu", "Fan Zhang", "Hao Wang", "Liang Liu", "Zhulin Xie", "Lijun Yao", "Xiao Liang", "Liangmin Wen", "Liqiang Feng", "Feilong Wu", "Min Hu", "Min Chen", "Guanjing Xiong", "Xiaohu Ruan", "Xiaoxin Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc15573bca49d453001ec22061f190e429704a5318751b041550e3610c469f4e", "submitted_at": "2026-09-11T03:36:35Z", "updated_at": "2026-09-15T06:36:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.12394v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.12394", "identifiers": {"arxiv": "2609.12394"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.12394"}], "name": "BlueLM-GUI Technical Report: A Real-Device-Centric Flywheel for Self-Improving Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a mobile-agent training pipeline centered on real-device rollouts, recovery of useful supervision from failures, and evolving evaluation tasks.", "topics": ["mobile-navigation", "reinforcement-learning", "data-generation"], "updated_at": "2026-09-15", "url": "https://arxiv.org/abs/2609.12394"}
{"aliases": [], "authors": ["Yu Bai", "Yukai Miao", "Dawei Wang", "Li Chen", "Yanyu Ren", "Yuqian Shi", "Dan Li", "Ying Xiong", "Chengqiu Tan", "Run Zhou", "Li Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff3b4cec8066a54cd58a6d4dd2b245536fa36a3157622565dc0b69ae7eb151f6", "submitted_at": "2026-09-11T03:54:02Z", "updated_at": "2026-09-11T03:54:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.12404v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.12404", "identifiers": {"arxiv": "2609.12404"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.12404"}], "name": "VRL-Bench: Benchmarking agents on computer control tasks under finite trial budgets", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-11", "url": "https://arxiv.org/abs/2609.12404"}
{"aliases": [], "authors": ["Zhutao Lv", "Chenhao Dang", "Yi Feng", "Yanpei Gong", "Xiaolei Wang", "Junyan Ye", "Conghui He", "Weijia Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f522778cfb2925fbadd6e0225651c24dd53c501bb0c8c423719222fb435ea255", "submitted_at": "2026-09-11T07:41:36Z", "updated_at": "2026-09-11T07:41:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.12533v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.12533", "identifiers": {"arxiv": "2609.12533"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.12533"}], "name": "Earth-Agent-Pro: Towards Real-World Full-Chain Earth Observation with Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-11", "url": "https://arxiv.org/abs/2609.12533"}
{"aliases": [], "authors": ["Zhangxuan Gu", "Haoxing Chen", "Qi Qin", "Yi Xin", "Kai Gan", "Lin Liu", "Long Cui", "Xiaomei Wang", "Beitong Zhou", "Yunzhu Zhang", "Zhengwen Zeng", "Changlong Gao", "Weizhi Chen", "Rongchao Zhang", "Haoyuan Wu", "Shuheng Shen", "Changhua Meng", "Weiqiang Wang", "Jianguo Li", "Zhenzhong Lan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d7e9643c0543f088e5d2b17699c843a7643c611315f20ce92e83aea849b931a", "submitted_at": "2026-09-09T05:53:39Z", "updated_at": "2026-09-09T05:53:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13287v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.13287", "identifiers": {"arxiv": "2609.13287"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.13287"}], "name": "LLaDA-UI: Bringing Block-wise Diffusion to Vision-Language GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "LLaDA-UI adapts block-wise diffusion generation to a vision-language GUI agent through multimodal pretraining and GUI supervised fine-tuning. It evaluates grounding and navigation across multiple platform settings.", "topics": ["grounding", "imitation-learning", "efficiency", "screen-understanding"], "updated_at": "2026-09-09", "url": "https://arxiv.org/abs/2609.13287"}
{"aliases": [], "authors": ["Alexandru Ianta", "Eleni Stroulia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29b615736812b466b0df7133b7070184a1b7c8dacbe4c73fe32b878b654bb78d", "submitted_at": "2026-09-11T19:57:22Z", "updated_at": "2026-09-11T19:57:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13491v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.13491", "identifiers": {"arxiv": "2609.13491"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.13491"}], "name": "Token Efficient Task Execution via Application Behavior Modeling for Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "OdoBot builds a model of web-application behavior from successful task demonstrations and uses it to guide task execution. Its Canvas LMS evaluation compares token use and completion with existing web agents.", "topics": ["memory", "efficiency", "web-navigation"], "updated_at": "2026-09-11", "url": "https://arxiv.org/abs/2609.13491"}
{"aliases": [], "authors": ["Xinyun Cao", "Adriana Szekeres", "Fazle Elahi Faisal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ce929c347ed5160b1f8455cc9d7e73ac2cc1b678e9d386d00dd4b2ef146b56d", "submitted_at": "2026-09-11T21:24:40Z", "updated_at": "2026-09-11T21:24:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13548v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.13548", "identifiers": {"arxiv": "2609.13548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.13548"}], "name": "AutoTailor: Automatic, User-Aligned Capability Selection and Adaptation for Web Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "AutoTailor compiles web trajectories into reusable browser-automation APIs, then filters and dynamically revises the capability inventory. WebArena experiments distinguish API-only operation from an agent that can fall back to ordinary reasoning and browser actions.", "topics": ["tool-integration", "web-navigation", "workflow", "efficiency"], "updated_at": "2026-09-11", "url": "https://arxiv.org/abs/2609.13548"}
{"aliases": [], "authors": ["Qihang Cen", "Tianshuo Cong", "Da Song", "Xinlei He", "Jiaxing Song", "Ke Xu", "Qi Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1eda33b358d4c4caf8c1fb0b1bd92698c3b37cee995d820f40ecffe82010edb4", "submitted_at": "2026-09-12T10:55:54Z", "updated_at": "2026-09-12T10:55:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13873v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.13873", "identifiers": {"arxiv": "2609.13873"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.13873"}], "name": "PriMobiBench: Characterizing Visual Privacy Leakage in VLM-Driven Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "PriMobiBench evaluates direct privacy leakage and user profiling from mobile GUI screenshot trajectories, supported by the MobiLeak trace dataset. It also tests masking sensitive, task-irrelevant interface elements before cloud processing and measures the resulting privacy and utility trade-off.", "topics": ["evaluation", "safety-security", "screen-understanding", "mobile-navigation"], "updated_at": "2026-09-12", "url": "https://arxiv.org/abs/2609.13873"}
{"aliases": [], "authors": ["Nitish Kovuru", "Prateek Jannu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c397c67c00690b079786b9481086be997d192497a8eff2c823d43c5ecbaf7d7", "submitted_at": "2026-09-13T02:22:37Z", "updated_at": "2026-09-13T02:22:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.14239v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.14239", "identifiers": {"arxiv": "2609.14239"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.14239"}], "name": "CoArena: Evaluating Computer-Use and Multi-Agent Systems in Real Time", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-13", "url": "https://arxiv.org/abs/2609.14239"}
{"aliases": [], "authors": ["Gongbo Zhang", "Hao Li", "Yu Wang", "Mujie Lin", "Liuzhenghao Lv", "Yicheng Mao", "Yimi Wang", "Jun Zhu", "Minhan Tang", "Zhengxiang Jiang", "Yusong Wang", "Jiayu Yao", "Kunpeng Ning", "Dawei Pang", "Yonghong Tian", "OpenAI4S Community", "Yuyang Liu", "Li Yuan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9eb6900955e5091439452ba4e07fb47d1ef2ddc88d862e63808193332a433934", "submitted_at": "2026-09-14T06:16:59Z", "updated_at": "2026-09-14T06:16:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.15096v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.15096", "identifiers": {"arxiv": "2609.15096"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.15096"}], "name": "OpenAI4S: Code as Action, Science as Sessions", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-14", "url": "https://arxiv.org/abs/2609.15096"}
{"aliases": [], "authors": ["Yunhao Feng", "Ruixiao Lin", "Ming Wen", "Yanming Guo", "Xingjun Ma", "Yutao Wu", "Xinhao Deng", "Shouling Ji"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4dd204d4efcb620479f444d4551cfb0037cff0dba0cc7c40ef4fc50b39376d10", "submitted_at": "2026-09-14T07:09:33Z", "updated_at": "2026-09-14T07:09:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.15134v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.15134", "identifiers": {"arxiv": "2609.15134"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.15134"}], "name": "HazardAuditor: From Executable Threats to Safer Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-14", "url": "https://arxiv.org/abs/2609.15134"}
{"aliases": [], "authors": ["Yizhou Liu", "Fei Tang", "Yuchen Yan", "Zhengxi Lu", "Songqin Nong", "Tao Jiang", "Wenhao Xu", "Wenqi Zhang", "Weiming Lu", "Jun Xiao", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e2081a382ae815b48eb3166d18ee7c6e53afb4cd6d75538f3ff3e29f2ed5954", "submitted_at": "2026-09-14T09:56:10Z", "updated_at": "2026-09-14T09:56:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.15307v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.15307", "identifiers": {"arxiv": "2609.15307"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.15307"}], "name": "Learning from Reliable Negatives: Confidence-Anchored Test-Time Adaptation for GUI Grounding", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Confidence-Anchored Learning uses coordinate-token confidence to filter pseudo-supervision for GUI grounding. Its negative-only variant performs label-free test-time updates to reduce reliance on potentially incorrect positive examples.", "topics": ["grounding", "robustness", "evaluation"], "updated_at": "2026-09-14", "url": "https://arxiv.org/abs/2609.15307"}
{"aliases": [], "authors": ["Sibo Zhu", "Shicheng Fan", "Xinyue Wang", "Wenyi Wu", "Kun Zhou", "Biwei Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c6ae88afe647542ffcd6715fa21ba7027a834fbb4399ea2107b014227783b5b5", "submitted_at": "2026-09-14T10:46:40Z", "updated_at": "2026-09-18T21:46:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.15364v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.15364", "identifiers": {"arxiv": "2609.15364"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.15364"}], "name": "RSIAgent: Autonomous Exploration for Recursive Self-improvement in New Environments", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-18", "url": "https://arxiv.org/abs/2609.15364"}
{"aliases": [], "authors": ["Shengjie Jin", "Zelong Sun", "Hengbo Xu", "Yanbiao Ma", "Zhiwu Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8a98337941972fadb725906b9d1315a0ca5921859bd04780cacbabb0500e6a5", "submitted_at": "2026-09-14T12:22:26Z", "updated_at": "2026-09-14T12:22:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.15457v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.15457", "identifiers": {"arxiv": "2609.15457"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.15457"}], "name": "AnchorGUI: Asymmetric Memory for Dual-Scale Learning in GUI Navigation", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "AnchorGUI compares expected and observed UI transitions and selectively retains screenshots when they disagree. This asymmetric memory supports immediate correction and focuses cross-attempt experience distillation on likely failure steps.", "topics": ["memory", "interaction-history", "state-tracking", "self-reflection", "efficiency"], "updated_at": "2026-09-14", "url": "https://arxiv.org/abs/2609.15457"}
{"aliases": [], "authors": ["Stephane Hatgis-Kessell", "W. Bradley Knox", "Emma Brunskill"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "850cdc7dd76067233f81e446152591e25fc49571512bdee90fc5b79873a72b1d", "submitted_at": "2026-09-14T13:28:52Z", "updated_at": "2026-09-14T13:28:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.15544v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.15544", "identifiers": {"arxiv": "2609.15544"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.15544"}], "name": "Specifying Reward Functions for RL Without Environment Sampling", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-14", "url": "https://arxiv.org/abs/2609.15544"}
{"aliases": [], "authors": ["Zihan Dong", "Yuanzhe Liu", "Zhiyuan Ma", "Qishi Zhan", "Dehan Kong", "Guohao Li", "Kaixin Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f5fd0a3024311f9aba44a2971c9c752ee2c38444139823746b90a7c9a6d5ed4", "submitted_at": "2026-09-14T19:19:05Z", "updated_at": "2026-09-21T14:30:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.16251v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.16251", "identifiers": {"arxiv": "2609.16251"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.16251"}], "name": "CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "CADWorld evaluates screenshot-and-GUI agents on long FreeCAD workflows, from modeling and assembly to manufacturing and simulation. Executable checks inspect saved engineering artifacts for geometry, constraints, construction structure and downstream state.", "topics": ["evaluation", "long-horizon", "workflow", "science"], "updated_at": "2026-09-21", "url": "https://arxiv.org/abs/2609.16251"}
{"aliases": [], "authors": ["Yao Zhao", "Aditya Shanmugham", "Swastik Roy", "Yanxun Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8dcd34beebb8a4660d8d7b9f4c76c3f86069926b9856d26b1c20d87dec83bb5", "submitted_at": "2026-09-15T04:56:17Z", "updated_at": "2026-09-15T04:56:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.16635v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.16635", "identifiers": {"arxiv": "2609.16635"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.16635"}], "name": "EchoPath: Execution-Level Replayable Memory for GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "EchoPath turns validated GUI trajectories into parameterized replay procedures with explicit runtime preconditions. Image-based target matching adjusts stored coordinates, while ambiguous steps fall back to bounded repair or fresh planning.", "topics": ["memory", "workflow", "automation", "grounding", "efficiency", "robustness"], "updated_at": "2026-09-15", "url": "https://arxiv.org/abs/2609.16635"}
{"aliases": [], "authors": ["Heng Li", "Fulin Zhao", "Zhe Geng", "Zhiyuan Yao", "Wei Yuan", "Xiapu Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b86516b7695d23ffc9179c0829209dc28b96c3c819525eab62a293a68ab5daf0", "submitted_at": "2026-09-15T07:02:10Z", "updated_at": "2026-09-15T07:02:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.16732v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.16732", "identifiers": {"arxiv": "2609.16732"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.16732"}], "name": "When Agents See Differently: Exposing UI Desynchronization Threats in Mobile Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Investigates mismatches between what humans see on a mobile display and what agents receive through screenshots or accessibility metadata. An automated evaluation framework embeds instruction-agnostic perturbations into functional APK clones and compares static and dynamic agent deviations with human noticeability.", "topics": ["safety-security", "screen-understanding", "evaluation", "robustness"], "updated_at": "2026-09-15", "url": "https://arxiv.org/abs/2609.16732"}
{"aliases": [], "authors": ["ZhuoXin Liu", "Zhiming Ma", "Ying Zhang", "Mengzheng Yang", "Yifan Wang", "Zhengqi Huang", "Yanhan Zhou", "Zekun Lin", "Jun Zhang", "Shun Zhang", "Yue Chen", "Qiao Zhao", "Peng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af3bdd004bf4ca0b8a0dfcea61c987b7ac958701f3ceae0159184fd78032fb68", "submitted_at": "2026-09-15T09:32:45Z", "updated_at": "2026-09-17T16:55:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.16900v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.16900", "identifiers": {"arxiv": "2609.16900"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.16900"}], "name": "RiskChainBench: A Benchmark for Obfuscated Platform Message Restoration and Evidence-Grounded Web Investigation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-17", "url": "https://arxiv.org/abs/2609.16900"}
{"aliases": [], "authors": ["Bofan Chen", "Boxuan Zhang", "Fei Tang", "Zhengxi Lu", "Yong Du", "Tongbo Chen", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "975662f216cc9a6b89856d4457f2125b733620e80bfac1b6424775253071d2e6", "submitted_at": "2026-09-15T17:24:06Z", "updated_at": "2026-09-15T17:24:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.17653v1"}}, "environments": ["android", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.17653", "identifiers": {"arxiv": "2609.17653"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.17653"}], "name": "Reflect, Revise, Reuse: Training-Free Skill Evolution for GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "EvoSkill-GUI revises reusable skill files from execution feedback through an executor–critic loop. It studies how training-free updates to localization, recovery and procedure information help subsequent mobile and desktop tasks.", "topics": ["memory", "training-free", "self-reflection", "robustness", "long-horizon", "workflow"], "updated_at": "2026-09-15", "url": "https://arxiv.org/abs/2609.17653"}
{"aliases": [], "authors": ["Kratika Bhagtani", "Kusha Sridhar", "Maziyar Baran Pouyan", "Yuying Zhao", "Eugene Siow"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ab0dbea71d0b8bfe035d3b0a71d6ed99fbd261202976cdc322a0769074827d76", "submitted_at": "2026-09-15T22:23:17Z", "updated_at": "2026-09-23T01:31:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.17885v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.17885", "identifiers": {"arxiv": "2609.17885"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.17885"}], "name": "ERPBench: A State-Grounded Evaluation Paradigm for Computer-Use Agents in Enterprise Software", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "ERPBench evaluates screenshot-only agents in enterprise software using database values to check the business records produced by their actions. It pairs these state-based checks with a harness for human approval and analyzes cases where visible completion masks incorrect persisted data.", "topics": ["evaluation", "workflow", "human-in-the-loop", "robustness"], "updated_at": "2026-09-23", "url": "https://arxiv.org/abs/2609.17885"}
{"aliases": [], "authors": ["Hasnain Irshad", "Anam Mughees", "Neelam Mughees", "Abdullah Mughees", "Imtiaz Ali Soomro"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0fcc5f2dde0b009516d13f055f6d43d4bd42deeb8d2f4751990a754a89a492d", "submitted_at": "2026-09-16T10:06:24Z", "updated_at": "2026-09-30T14:38:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.18411v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.18411", "identifiers": {"arxiv": "2609.18411"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.18411"}], "name": "The Verifiable Action Card: Trustworthy Human-in-the-Loop Control for Secure Autonomous Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2609.18411"}
{"aliases": [], "authors": ["Liyang Fan", "Xinping Bi", "Yitai Li", "Shuaimin Li", "Hui Li", "Min Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccbc85f4d2598b4cb1e6e1a96b3b9184ddeba838cb610658070fd2d1ac9c8d06", "submitted_at": "2026-09-16T14:06:38Z", "updated_at": "2026-09-16T14:06:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.18690v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.18690", "identifiers": {"arxiv": "2609.18690"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.18690"}], "name": "RankGround: Efficient High-Resolution GUI Grounding via Lightweight Reranker-Guided Crop Selection", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "RankGround trains a lightweight reranker to select a promising screenshot crop before one VLM localization call. Ranking supervision uses containment and boundary-aware augmentation, with pointwise and listwise training stages.", "topics": ["grounding", "efficiency", "data-generation", "screen-understanding"], "updated_at": "2026-09-16", "url": "https://arxiv.org/abs/2609.18690"}
{"aliases": [], "authors": ["Jin Gao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5744036c81b58b4e49cf52001390e1e2aa5341599b30faac09afa755ba1b14b", "submitted_at": "2026-09-16T17:46:34Z", "updated_at": "2026-09-16T17:46:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.19125v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.19125", "identifiers": {"arxiv": "2609.19125"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.19125"}], "name": "Affora: A Design System for Agent-Friendly Interfaces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-16", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-16", "url": "https://arxiv.org/abs/2609.19125"}
{"aliases": [], "authors": ["Yinzhu Quan", "Zefang Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e7fb5704932a99147ff7c64a82a6a4041565f3edf6506a6252f75f832834d96", "submitted_at": "2026-09-17T00:29:46Z", "updated_at": "2026-10-08T04:53:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.19523v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.19523", "identifiers": {"arxiv": "2609.19523"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.19523"}], "name": "EconSkills: Studying Skill Transfer and Retrieval for Web Agents on Live Economic Data", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "EconSkills converts verified economic-data browsing traces into parameterized procedures with navigation, verification and recovery guidance. Its evaluation separates known-skill transfer from library retrieval, exposing the limits of approximate matches to new tasks.", "topics": ["memory", "web-navigation", "evaluation", "workflow"], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2609.19523"}
{"aliases": [], "authors": ["Haya Halimeh", "Sascha Kaltenpoth", "Kevin Bösch", "Oliver Müller"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "70a7588abcf345418ed759fb2bdb96fc60c6e1444a3e0dc98766510c755f8478", "submitted_at": "2026-09-17T07:54:15Z", "updated_at": "2026-09-17T07:54:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.19843v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.19843", "identifiers": {"arxiv": "2609.19843"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.19843"}], "name": "A Dual-Process Perspective on Nudge Susceptibility in LLM-Based GUI Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies how automatic defaults and reflective social-influence cues affect GUI-agent shopping decisions under different reasoning configurations. A randomized simulated-shopping experiment finds that reasoning changes the type of susceptibility rather than uniformly removing interface influence.", "topics": ["safety-security", "evaluation", "ui-design", "robustness"], "updated_at": "2026-09-17", "url": "https://arxiv.org/abs/2609.19843"}
{"aliases": [], "authors": ["Zofia Smoleń"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9d2d9616e957c0dbce9bdb4b913a38ec9761aa630d3dea320d541f477db35c7", "submitted_at": "2026-09-17T17:22:43Z", "updated_at": "2026-09-17T17:22:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.20732v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.20732", "identifiers": {"arxiv": "2609.20732"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.20732"}], "name": "Q&A on Any Spreadsheet Requires Interpreting Its Grid Structure", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-17", "url": "https://arxiv.org/abs/2609.20732"}
{"aliases": [], "authors": ["Xiuhui Zhang", "Yi Chen", "Shusheng Xu", "Fan Li", "Huan Wang", "Tongkai Yang", "Binhang Yuan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4e68c2e7eab16d1295bd37b6a2264e4424486eee7d6dfedc3bac0d0675e1d37b", "submitted_at": "2026-09-18T04:06:55Z", "updated_at": "2026-09-18T04:06:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.21293v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.21293", "identifiers": {"arxiv": "2609.21293"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.21293"}], "name": "GameASG-Bench: Benchmarking Autonomous Software Generation for Game Development", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-18", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-18", "url": "https://arxiv.org/abs/2609.21293"}
{"aliases": [], "authors": ["Shuai Bai", "Jiayong Deng", "Sicheng Fan", "Yikun Fu", "Chang Gao", "Xuhao Hu", "Mianqiu Huang", "Yizhen Jiang", "Yuheng Jing", "Dehui Kong", "Keliang Li", "Ning Li", "Wanli Li", "Dayiheng Liu", "Dunjie Lu", "Changwei Luo", "Que Shen", "Zheyuan Wang", "Zijian Wang", "Jie Wu", "Gao Wu", "Zhihui Xie", "Rui Xie", "Haiyang Xu", "An Yang", "Jiakang Yuan", "Yanming Zhang", "Jiajun Zhang", "Xi Zhang", "Zhenru Zhang", "Zhuo Zhen", "Mingkang Zhu", "Bowen Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3b312c5a4d638ec9f9a7408fd9ce2d8cbf74fa462d48c8aa37f59e4ad3bd8edc", "submitted_at": "2026-09-18T17:00:56Z", "updated_at": "2026-09-21T07:09:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.22000v2"}}, "environments": ["desktop", "linux", "macos", "windows", "android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.22000", "identifiers": {"arxiv": "2609.22000"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.22000"}], "name": "RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "RecreationWorld gives hybrid agents running reference applications to explore and recreate using native GUI controls and coding tools. Its multi-platform benchmark checks interaction-dependent behavior with programmatic and visual assertions, while trajectory-training transfer is evaluated separately.", "topics": ["evaluation", "data-generation", "tool-integration", "workflow"], "updated_at": "2026-09-21", "url": "https://arxiv.org/abs/2609.22000"}
{"aliases": [], "authors": ["Mingfei Gao", "Rui Tian", "Haiming Gang", "Bohan Zhai", "Le Zhang", "Yuanzheng Gong", "Di Feng", "Ege Özsoy", "Kaixin Ma", "Vishwesh Kirthivasan", "Oğuzhan Fatih Kar", "Roman Bachmann", "Anders Boesen Lindbo Larsen", "Afshin Dehghan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4dfb0703133032426ff8f40ed6a2c8131025b07658c0b643603f8d50ba163d16", "submitted_at": "2026-09-18T17:59:34Z", "updated_at": "2026-09-18T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.22083v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.22083", "identifiers": {"arxiv": "2609.22083"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.22083"}], "name": "MintAct: A Unified Visual Agent for Digital Environments", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-18", "relations": [], "review_status": "curated", "scope": "core", "summary": "MintAct trains a visual-agent family for grounding, navigation and tool use across mobile, desktop and web environments. Its shared infrastructure supports trajectory collection and asynchronous online RL while controlling cross-domain sampling and off-policy drift.", "topics": ["grounding", "reinforcement-learning", "data-generation", "tool-integration"], "updated_at": "2026-09-18", "url": "https://arxiv.org/abs/2609.22083"}
{"aliases": [], "authors": ["Zeyu Kang", "Zhenyun Yin", "Yang Zhang", "Shan He", "Shanzhe Lei", "Yanjiu Zhong", "Xinquan Chen", "Xuhong Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a88d287d5433009962a905dbdbf03359e1183e66e4a6015441ec28405f64a6ee", "submitted_at": "2026-08-27T05:43:50Z", "updated_at": "2026-09-22T06:39:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.22178v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.22178", "identifiers": {"arxiv": "2609.22178"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.22178"}], "name": "Beyond Task Completion: Training Capable and Safe Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SCOPE, which trains computer-use agents on ordinary task demonstrations, safe continuations around hazards, and explicit refusals. Its SCOPE-Gen pipeline pairs verifiable tasks with environmental-risk variants to support joint capability and safety training.", "topics": ["safety-security", "reinforcement-learning", "data-generation"], "updated_at": "2026-09-22", "url": "https://arxiv.org/abs/2609.22178"}
{"aliases": [], "authors": ["Changyue Jiang", "Jiayi Wang", "Xin Wen", "Jiarun Dai", "Geng Hong", "Xudong Pan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d43e0803b924cacdc9b7af29757ed2539b6f3e09bed3c263d6465bbfc550733", "submitted_at": "2026-09-19T03:18:25Z", "updated_at": "2026-09-19T03:18:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.22724v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.22724", "identifiers": {"arxiv": "2609.22724"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.22724"}], "name": "MATE: Policy-Aware Security Auditing for Mobile Agents via Synthesis-Driven Trajectory Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-19", "url": "https://arxiv.org/abs/2609.22724"}
{"aliases": [], "authors": ["Dongming Jiang", "Yi Li", "Bingzhe Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c66fb393bd6a5a64e1fa3724c6239bc51dc0abe69ee8aa1ad5511078481d9532", "submitted_at": "2026-09-21T01:43:12Z", "updated_at": "2026-09-21T01:43:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.23986v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.23986", "identifiers": {"arxiv": "2609.23986"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.23986"}], "name": "Jev-Mem: System-One-Controlled Agentic Memory for Efficient AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-21", "url": "https://arxiv.org/abs/2609.23986"}
{"aliases": [], "authors": ["Yunxiang Li", "Xixin Wu", "Helen Meng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64936af65d0ca351b19da3f63325f2bc2a5c9e0951e5a7806cd4dd82b18784a7", "submitted_at": "2026-09-21T08:42:08Z", "updated_at": "2026-09-21T08:42:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.24277v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.24277", "identifiers": {"arxiv": "2609.24277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.24277"}], "name": "How Many Pixels Is a Digit Worth? Place-Aware Coordinate Entropy for GUI Agent Confidence Estimation", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "PACE weights coordinate-digit entropy by place value to estimate confidence in GUI clicks from one forward pass. It evaluates correctness ranking and selective accuracy against repeated-sampling alternatives on ScreenSpot-Pro and ScreenSpot-v2.", "topics": ["grounding", "efficiency", "evaluation", "robustness"], "updated_at": "2026-09-21", "url": "https://arxiv.org/abs/2609.24277"}
{"aliases": [], "authors": ["Zhilin Wang", "Shaokun Zhang", "Yifan Zhang", "Hao Zhang", "Jin Xu", "Binfeng Xu", "Jian Hu", "Yunheng Zou", "Karan Sapra", "Andrew Tao", "Jan Kautz", "Yi Dong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d3cb920a677108910530788eb68c263127dba9510274f182910aed9c2f848c9", "submitted_at": "2026-09-21T16:55:24Z", "updated_at": "2026-09-21T16:55:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.24890v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.24890", "identifiers": {"arxiv": "2609.24890"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.24890"}], "name": "OSWorld-Pro: Process-based Evaluation for Computer Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-21", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adds annotated intermediate subgoals to computer-use tasks and uses process-oriented evaluation to identify action and progress failures.", "topics": ["evaluation", "observability"], "updated_at": "2026-09-21", "url": "https://arxiv.org/abs/2609.24890"}
{"aliases": [], "authors": ["Soumil Rathi", "Deshraj Yadav", "Taranjeet Singh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "839c584110c90641b6ff3a360f3d1616637d43020916cd08992c6e9285c184dc", "submitted_at": "2026-09-21T17:54:35Z", "updated_at": "2026-09-22T06:04:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.24971v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.24971", "identifiers": {"arxiv": "2609.24971"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.24971"}], "name": "DolphinBench: Mapping the Pareto Frontier of Agent Memory", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-22", "url": "https://arxiv.org/abs/2609.24971"}
{"aliases": [], "authors": ["Binxu Li", "Haoyi Duan", "Yuhui Zhang", "Yaohui Zhang", "Zihao Lin", "Kaituo Feng", "Suozhi Huang", "Xiangyi Li", "Yu Li", "Chunyuan Li", "Shilong Liu", "Mengdi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43007f4c10270fb742ea912690f607874e2f2ea2cf6e025df2500e16cf8d0116", "submitted_at": "2026-09-21T17:58:56Z", "updated_at": "2026-10-05T16:40:27Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.24997v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.24997", "identifiers": {"arxiv": "2609.24997"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.24997"}], "name": "VideoGen-Agent: Reinforcing Video Generation Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2609.24997"}
{"aliases": [], "authors": ["Yan Zhang", "Pei Fu", "Daiqing Wu", "Huawen Shen", "Ruoceng Zhang", "Shaojie Zhang", "Jiahui Yang", "Yu Zhou", "Can Ma", "Zhenbo Luo", "Jian Luan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d184b111aa2a9ad65dd57391b72bcef369c2bcd01ceb8e4f00f7a2320db856d8", "submitted_at": "2026-09-22T07:02:35Z", "updated_at": "2026-09-22T07:02:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.25769v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.25769", "identifiers": {"arxiv": "2609.25769"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.25769"}], "name": "Towards Omni-dimensional GUI Agent Navigation with Masked Trajectory Prediction", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-22", "relations": [], "review_status": "curated", "scope": "core", "summary": "MaP unifies GUI-navigation objectives by masking and predicting components of multi-turn trajectories. Role-aware adapters address heterogeneous training data, and evaluations compare the approach with direct task-mixture training across navigation benchmarks.", "topics": ["imitation-learning", "action-prediction", "state-tracking", "planning"], "updated_at": "2026-09-22", "url": "https://arxiv.org/abs/2609.25769"}
{"aliases": [], "authors": ["Yuxuan Li", "Will Epperson", "Wesley Deng", "Zezhou Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8d241c1ea9731bc815eef5df9c409d410608adb052616406d955d7d33cecd9b9", "submitted_at": "2026-09-23T03:00:00Z", "updated_at": "2026-09-27T20:49:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.27273v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.27273", "identifiers": {"arxiv": "2609.27273"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27273"}], "name": "CAVEAT: Towards Robust Computer-Use Agents in Incentive-Misaligned Environments", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CAVEAT, a controlled marketplace benchmark for testing whether delegated agents preserve user objectives when interfaces favor other outcomes. It diagnoses decision failures and studies harness interventions and targeted training to improve resistance to marketplace steering.", "topics": ["evaluation", "robustness", "safety-security"], "updated_at": "2026-09-27", "url": "https://arxiv.org/abs/2609.27273"}
{"aliases": [], "authors": ["Yan Zhang", "Daiqing Wu", "Huawen Shen", "Liang Li", "Gang Cao", "Zhi Gong", "Wei Dai", "Xiaode Zhang", "Can Ma", "Yu Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ae6c83c017a05c2916f31e388f6584cc2cab7b87e1eb0038c33a707bbd66e24", "submitted_at": "2026-09-23T03:38:11Z", "updated_at": "2026-09-23T03:38:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27307v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.27307", "identifiers": {"arxiv": "2609.27307"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27307"}], "name": "Learn How to Act from Your Own Interactions: On-Policy Self-Distillation for GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-SD-v2 extends on-policy self-distillation from grounding to multi-turn GUI interaction. It first strengthens responses to privileged guidance, then selectively distills step-specific reasoning and memory support for later actions.", "topics": ["imitation-learning", "memory", "interaction-history", "mobile-navigation"], "updated_at": "2026-09-23", "url": "https://arxiv.org/abs/2609.27307"}
{"aliases": [], "authors": ["Chengguang Gan", "Yunhao Liang", "QingHao Zhang", "Shiwen Ni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b2bda1c50fd47fabbe77c40a2938979a6a16b407b40f11737fc6f1fc795d090", "submitted_at": "2026-09-23T04:49:07Z", "updated_at": "2026-09-23T04:49:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27353v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.27353", "identifiers": {"arxiv": "2609.27353"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27353"}], "name": "Guides That Cause Actions: An Offline Study of Guide-Action Mutual Reinforcement in Multimodal Web Agents", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebMRE pairs guidance sentences with grounded actions in an offline benchmark derived from successful WebArena traces. Controlled decoding-prefix interventions test how guide content affects action selection, while separating offline prediction metrics from live browser task success.", "topics": ["evaluation", "action-prediction", "imitation-learning", "grounding"], "updated_at": "2026-09-23", "url": "https://arxiv.org/abs/2609.27353"}
{"aliases": [], "authors": ["Xueji Fang", "Boqiang Duan", "Hua Wu", "Jingdong Wang", "Guo-Jun Qi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e820cb77813fdb1059c18fa49bb134b3984af37b4e8a7e78284722206f3552af", "submitted_at": "2026-09-23T07:22:43Z", "updated_at": "2026-10-10T05:28:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.27455v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.27455", "identifiers": {"arxiv": "2609.27455"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.27455"}], "name": "LeWAM: Latent evolving World Action Model", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-23", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-10", "url": "https://arxiv.org/abs/2609.27455"}
{"aliases": [], "authors": ["Zeyu Zhang", "Dingyi Rong", "Zijian Chen", "Zicheng Zhang", "Xiongkuo Min", "Guangtao Zhai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a89489bb1ff3001a1d806c3eefd8a6aa19503d1dfbe0265b916766119f96ec2", "submitted_at": "2026-09-23T07:27:23Z", "updated_at": "2026-09-23T07:27:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27461v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.27461", "identifiers": {"arxiv": "2609.27461"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27461"}], "name": "Invisible in Space, Visible in Time: Motion Vision CAPTCHA against GUI Agents", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-23", "relations": [], "review_status": "curated", "scope": "core", "summary": "MVCAP-Bench tests live browser CAPTCHAs whose foreground meaning emerges through motion against dynamic backgrounds. A foreground-only control separates temporal camouflage difficulty from answer format and browser interaction.", "topics": ["evaluation", "screen-understanding", "video", "safety-security"], "updated_at": "2026-09-23", "url": "https://arxiv.org/abs/2609.27461"}
{"aliases": [], "authors": ["Zhitong Guo", "Jing Yu Koh", "Ruiyu Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "80f6fb973e9f8d62e515d02105d08be774fc32ed5b20a7ecee92b45d112415aa", "submitted_at": "2026-08-16T19:54:53Z", "updated_at": "2026-08-16T19:54:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27770v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.27770", "identifiers": {"arxiv": "2609.27770"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27770"}], "name": "Improving LLM-based Autonomous Web Agents with Filtering", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-08-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains HTML-element relevance filters on Mind2Web trajectories and transfers them to WebArena to reduce irrelevant browser context. It also evaluates a zero-shot ColBERT retriever and documents baseline web-agent failure modes.", "topics": ["web-navigation", "screen-understanding", "efficiency", "evaluation"], "updated_at": "2026-08-16", "url": "https://arxiv.org/abs/2609.27770"}
{"aliases": [], "authors": ["Zhenhua Zou", "Sheng Guo", "Qiuyang Zhan", "Lepeng Zhao", "Shuo Li", "Zhuotao Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e273628ed6a72b5ed492e8cc159f26b91e0874f7be7e6db4f4d781577a9604a3", "submitted_at": "2026-08-29T12:20:17Z", "updated_at": "2026-08-29T12:20:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.29647v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.29647", "identifiers": {"arxiv": "2609.29647"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.29647"}], "name": "AgentKernel: The Trust-Native Agentic Operating System", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-08-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-08-29", "url": "https://arxiv.org/abs/2609.29647"}
{"aliases": [], "authors": ["Tingyu Qu", "Weigao Sun", "Yuecheng Liu", "Yucheng Zhao", "Yi Zhu", "Yifeng Ding", "Qiyi Wang", "Sihan Cao", "Pengkun Jiao", "Hanlei Xie", "Xiongwei Wu", "Qichao Wang", "Haodong Zhang", "Jiajun Liu", "Yuhao Wang", "Yuqing Xie", "Junpeng Zhao", "Long Chen", "Ming Ma", "Sihan Yang", "Ziwang Zhao", "Yanhao Jia", "Liangquan Gong", "Feida Zhu", "Yiran Zhong", "Steven Hoi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "572ba5397fc1e4bb138febc3e7004e6e4f244e0ba8b628fb91debef4a854561f", "submitted_at": "2026-09-24T14:38:08Z", "updated_at": "2026-09-24T14:38:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.29892v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2609.29892v1"], "reason": "Record the paper-specific arXiv document declaration; it does not establish code, data or model-weight permissions."}}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:43:20.402537Z", "role": "primary", "sha256": "d4385231cea2965a48895eeebd2dd6000265dd7b43731a0ad88b0373a21d0308", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2609.29892v1"}], "id": "arxiv:2609.29892", "identifiers": {"arxiv": "2609.29892"}, "kind": "paper", "last_verified_at": "2026-10-20T15:43:20.402537Z", "licenses": {"code": null, "data": null, "document": {"basis": "The exact-v1 arXiv HTML infobox (#license-tr) declares License: CC BY 4.0 for this paper. This declaration is document-only.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T15:43:20.402537Z", "url": "https://arxiv.org/html/2609.29892v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.29892"}], "name": "Qwen-Planner-Agent: A Closed-Loop AI-for-AI Framework for Real-World Mobile Planner Agents", "notes": "Primarily typed tool actions and structured feedback, not pixel-coordinate GUI control; backends combine programmatic sandboxes, LLM simulation and selected real-device sessions. No specific OS is established. Model/harness updates are reviewed offline, not autonomous serving-time changes. Co-evolution scores use an evolve/development set; output-cost estimates exclude input, tools, devices and harness processing. Memory comparisons differ in history access and processing cost; user authorization relies on trusted upstream services. No independent reproduction.", "published_at": "2026-09-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains a mobile planner with supervised trajectories and competence-adaptive RL, then couples typed tool execution with memory, skills and offline model-harness refinement.", "topics": ["planning", "data-generation", "reinforcement-learning", "memory", "mobile-navigation", "efficiency"], "updated_at": "2026-09-24", "url": "https://arxiv.org/abs/2609.29892"}
{"aliases": [], "authors": ["Linghua Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b387c1cfe77f6e06477415944b3923189a1f0ca30be0680f01815cb65b0067bc", "submitted_at": "2026-09-24T17:30:32Z", "updated_at": "2026-09-24T17:30:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.30186v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.30186", "identifiers": {"arxiv": "2609.30186"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.30186"}], "name": "Jev-Mobile: Jev as an Executor for Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates occasional visual-language planning from frequent lightweight action selection over an accessibility-tree-based action space.", "topics": ["mobile-navigation", "efficiency", "planning"], "updated_at": "2026-09-24", "url": "https://arxiv.org/abs/2609.30186"}
{"aliases": [], "authors": ["Alexander Gill", "Md Farhan Ishmam", "Xuyen Nguyen", "Neha Bhat", "Parker Henry DeYoung", "Fateme Hashemi Chaleshtori", "Nathan Stringham", "Kenneth Marino", "Ana Marasović"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6ed206d8f8fec53c03956fffcafec6b55d94a47fc705bbc8b77c832b76419e6", "submitted_at": "2026-09-24T22:36:10Z", "updated_at": "2026-09-24T22:36:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.30604v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.30604", "identifiers": {"arxiv": "2609.30604"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.30604"}], "name": "The Hard Part Comes After Search: Benchmarking Web Agents on Synthesizing, Organizing, and Displaying Knowledge", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.", "published_at": "2026-09-24", "relations": [], "review_status": "curated", "scope": "core", "summary": "KNOWS benchmarks browser workflows that retrieve and synthesize information into documents, presentations and spreadsheets. Task-specific evaluators combine deterministic checks with model judgments, distinguishing partial progress from usable end-to-end artifact completion.", "topics": ["evaluation", "web-navigation", "workflow", "long-horizon"], "updated_at": "2026-09-24", "url": "https://arxiv.org/abs/2609.30604"}
{"aliases": [], "authors": ["Shivam Negi", "Arpit Rawat", "Rashi Jain"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "44bc2adddd2590051e363f425e4da8825ddecf490d8770402a70983e91912945", "submitted_at": "2026-09-25T04:22:20Z", "updated_at": "2026-09-25T04:22:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.30798v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.30798", "identifiers": {"arxiv": "2609.30798"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.30798"}], "name": "Evaluating Real-Time Voice Agents: From Component Quality to Grounded Outcomes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-25", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-25", "url": "https://arxiv.org/abs/2609.30798"}
{"aliases": [], "authors": ["Yuchen Sun", "Chenglin Cai", "Gongjie Zhang", "Tianyu Xia", "Quyu Kong", "Panrong Tong", "Zhengwen Zeng", "Long Chen", "Steven Hoi", "Chongyang Zhang", "Yue Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "838635f067b706c7b36a948c8100b30bf4cacd218470aaed6970e5186fa44fb9", "submitted_at": "2026-09-25T06:50:04Z", "updated_at": "2026-09-25T06:50:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.30887v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:30.162100+00:00", "role": "primary", "sha256": "096098b62f82867ee39dfd7cf44095edc5c6a9307351265592e485d82eadf111", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202609200000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T10:53:44.035673+00:00", "role": "primary", "sha256": "2b6b8147cf7b0e1fcded85f397c272bcf535ed179789e76c3ff4f15aa3fe53a7", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2609.30887v1"}], "id": "arxiv:2609.30887", "identifiers": {"arxiv": "2609.30887"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:30.162100+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.30887"}], "name": "From Tapping to Hopping: Augmenting Mobile GUI Agents with App-Native Deeplinks", "notes": "Reviewed the complete retained Atom abstract and selected exact-v1 full text. The Atom has incomplete clauses; its text/hash remain unchanged and this annotation uses the explicit methods. Evaluation is Android-only, with a selected real-app subset. Landing-Activity verification is not task completion. Model/catalog release is prospective in the paper. No independent reproduction.", "published_at": "2026-09-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-Hopper combines verified Android deeplinks with GUI actions, using landing-screen descriptions, hybrid trajectory training and reinforcement learning to select navigation shortcuts and recover through GUI control.", "topics": ["mobile-navigation", "tool-integration", "data-generation", "imitation-learning", "reinforcement-learning", "efficiency"], "updated_at": "2026-09-25", "url": "https://arxiv.org/abs/2609.30887"}
{"aliases": [], "authors": ["Kiana Hooshanfar", "Amirhossein Kazerouni", "Alireza Hosseini", "Michael Brudno", "Babak Taati"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ce76f34abe391289643321027a33ac3cac8b5bd64eaa901f9a5fbf7c5561362", "submitted_at": "2026-09-25T14:59:14Z", "updated_at": "2026-09-25T14:59:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.31364v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2609.31364v1"], "reason": "Document declaration only. The paper does not establish code, model-weight or underlying dataset permissions through this license."}}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T19:01:29.156954Z", "role": "primary", "sha256": "cd4ba28778b725c63f4a7329f683b2a287db2e46ed228a0d47006fcd95bbce80", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2609.31364v1"}], "id": "arxiv:2609.31364", "identifiers": {"arxiv": "2609.31364"}, "kind": "paper", "last_verified_at": "2026-10-20T19:01:29.156954Z", "licenses": {"code": null, "data": null, "document": {"basis": "The exact-v1 arXiv HTML infobox (#license-tr) declares License: CC BY 4.0 for this paper document.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T19:01:29.156954Z", "url": "https://arxiv.org/html/2609.31364v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.31364"}], "name": "OpenVAM: Open-World Visual Attention Modeling with VLMs", "notes": "Static human-attention modeling on U-EYE webpage screenshots, with FIWI held out from training; no live-browser, agent-action or task-success evaluation. Gemini-generated reference rationales are expert-checked, and JScore compares generated text with references; human validation covers a separate 100-image subset. Reference text and heatmap alignment do not establish causal explanations of gaze or temporal scanpaths. Several baseline scores come from prior papers. Code/data/weight releases and licenses were not verified; no independent reproduction.", "published_at": "2026-09-25", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Combines dense saliency prediction with explanations of attended elements, including webpage layouts; evaluates human-attention agreement and textual rationales rather than agent actions.", "topics": ["screen-understanding", "ui-design", "evaluation", "data-generation"], "updated_at": "2026-09-25", "url": "https://arxiv.org/abs/2609.31364"}
{"aliases": [], "authors": ["Zhenhao Zhang", "Zhaoyu Fan", "Haohan Ying", "Jingwen Hu", "Hancen Fan", "Junhao Zhou", "Zitian Chen", "Linchao Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa2bcfa05fcbd6937af6ba3d23e892e5a105248939538dcdc9d7b015daf38593", "submitted_at": "2026-09-25T20:04:24Z", "updated_at": "2026-09-25T20:04:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.31957v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.31957", "identifiers": {"arxiv": "2609.31957"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.31957"}], "name": "CaptchaArena: A Large-Scale, Fine-Grained Dataset for Training Computer-Use Agents on Interactive CAPTCHAs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CaptchaArena, a dataset of execution-verified interactive CAPTCHA solutions with screenshot–action trajectories, reasoning annotations, and pixel masks for irregular targets. The work trains a single policy across CAPTCHA types using supervised learning followed by verifier-reward reinforcement learning.", "topics": ["data-generation", "grounding", "reinforcement-learning"], "updated_at": "2026-09-25", "url": "https://arxiv.org/abs/2609.31957"}
{"aliases": [], "authors": ["Chenyue Li", "Xiaoxiao Sun", "Yubo Deng", "Qinlin Zhao", "Serena Yeung-Levy", "Yuhui Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "12976e58eb48d9c4709fb03508ac7ed7eeafffd3df675d3dccfd09bdc354f3db", "submitted_at": "2026-09-25T22:07:29Z", "updated_at": "2026-09-25T22:07:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32036v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.32036", "identifiers": {"arxiv": "2609.32036"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32036"}], "name": "ScreenHaystack: Finding Blind Zones in GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-25", "relations": [], "review_status": "curated", "scope": "core", "summary": "Moves controlled target icons across high-resolution interfaces to diagnose spatial blind zones in grounding models and study targeted augmentation.", "topics": ["grounding", "evaluation", "robustness"], "updated_at": "2026-09-25", "url": "https://arxiv.org/abs/2609.32036"}
{"aliases": [], "authors": ["Beining Wu", "Jun Huang", "Yanxiao Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "719385c97b9d174bbebc34ac4a196971af61521489f7bf032a7d807ae3c8d4c1", "submitted_at": "2026-09-26T01:13:44Z", "updated_at": "2026-09-26T01:13:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32130v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.32130", "identifiers": {"arxiv": "2609.32130"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32130"}], "name": "REBASE: Device-Cloud Experience Coherence for GUI Agents Across App Updates", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces REBASE, a device–cloud protocol for repairing cached GUI experience after app updates. It combines cloud replay, device-side step verification, and progressively richer failure reports to produce repairs tied to application versions.", "topics": ["memory", "robustness", "efficiency", "mobile-navigation"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32130"}
{"aliases": [], "authors": ["Taehwan Park", "Changmin Lee", "Hayeon Lee", "Taesik Gong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f343a941e12c0fb8c4e12dc156e7c8b37f753d74836d8e42542d504bee883e51", "submitted_at": "2026-09-26T02:41:56Z", "updated_at": "2026-09-26T02:41:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32166v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.32166", "identifiers": {"arxiv": "2609.32166"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32166"}], "name": "PastForward: Faster On-Device GUI Agents via Computational Experience Reuse", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PastForward, which reuses earlier output sequences as verified decoding proposals and starts future action computation from predicted GUI transitions. Screen checks determine whether speculative computation and cached model states can be retained.", "topics": ["memory", "efficiency", "mobile-navigation"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32166"}
{"aliases": [], "authors": ["Yangzhe Peng", "Xiaoyang Wang", "Yiyang Zhao", "Lijun Wu", "Kun He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cc6fd87d2299e2881eab9a2db96e4022663e2d94da5549dcde166c0c8b8be11", "submitted_at": "2026-09-26T08:14:50Z", "updated_at": "2026-09-26T08:14:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32351v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.32351", "identifiers": {"arxiv": "2609.32351"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32351"}], "name": "From Trajectories to Grounded Preferences: Process Preference Synthesis via Interaction Element Graphs for Web PRMs", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SURFPRM, which turns web demonstrations into an interaction-element graph for synthesizing grounded negative actions. Its contrastive preference data trains process reward models to distinguish plausible on-page alternatives and guide downstream trajectory search.", "topics": ["data-generation", "reward-models", "web-navigation"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32351"}
{"aliases": [], "authors": ["Chuhan Zhang", "Qi Xie", "Ziyue Wang", "Jianing Yin", "Yunfan Zhou", "Dazhen Deng", "Yingcai Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f091cab38c932e08d21afa936dfe241c93c0b2139aba1d1ce7a6d1513ad5da1a", "submitted_at": "2026-09-26T08:58:50Z", "updated_at": "2026-09-26T08:58:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32385v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.32385", "identifiers": {"arxiv": "2609.32385"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32385"}], "name": "DashAct: A Progressive Diagnostic Benchmark for GUI Agents in Interactive Dashboard Analysis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Diagnoses dashboard-agent failures by progressively restoring interaction context, next-action support, and visual target information within the same task.", "topics": ["evaluation", "grounding", "planning"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32385"}
{"aliases": [], "authors": ["Jinming Hu", "Haodong Zhao", "Qi Jia", "Die Chen", "Tianhang Zhao", "Sufeng Duan", "Gongshen Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff1c4eff95094ea8ecf37b7bc621eebbd5e4a677b0e11240e802a72b9b229c7f", "submitted_at": "2026-09-26T11:59:39Z", "updated_at": "2026-09-26T11:59:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32511v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:08:56Z", "role": "primary", "sha256": "ce7210dccecb2896202639e6f7be41876fee20de1135b0d23278cead2d695339", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2609.32511v1"}], "id": "arxiv:2609.32511", "identifiers": {"arxiv": "2609.32511"}, "kind": "paper", "last_verified_at": "2026-10-20T15:08:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32511"}], "name": "Learning from Others, Acting for You: Cross-User Memory Sharing for LLM Agents", "notes": "Main Mind2Web evaluation uses 252 recorded tasks assigned to 20 simulated users, frozen demonstration-derived memory and task-macro step metrics; it is not a live-browser completion test. The main split withholds same-site local experience, and sharing gains shrink when local coverage rises. Mind2Web has no preference channel; preference claims come from separate service and coding settings. User-bound retrieval is not a demonstrated privacy or security guarantee. Selected sections were source-read; no independent reproduction.", "published_at": "2026-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "ShareMem consolidates reusable experience across users while keeping preference retrieval user-bound; its web evaluation predicts target elements and actions on recorded Mind2Web states.", "topics": ["memory", "web-navigation", "action-prediction", "evaluation"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32511"}
{"aliases": [], "authors": ["Prince Zizhuang Wang", "Chenhao Liang", "Zelong Xu", "Aojie Yuan", "Xiaolin Zhou", "Haiyue Zhang", "Yue Zhao", "Xiyang Hu", "Shuli Jiang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d53f2512b99828849f047f63412264437a59b730c117c801de252e6dc80a33fc", "submitted_at": "2026-09-26T13:18:00Z", "updated_at": "2026-09-26T13:18:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32600v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.32600", "identifiers": {"arxiv": "2609.32600"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32600"}], "name": "CUA-SWE: When Computer-Use Agents Meet Visual Software Engineering", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CUA-SWE, a benchmark and execution environment that combines code changes with graphical interaction and visual inspection of running software. Task-specific tests verify repairs and changes whose required information may be available only through the application interface.", "topics": ["evaluation", "workflow", "tool-integration"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32600"}
{"aliases": [], "authors": ["Dongsheng Liu", "Chao Jin", "Wenkui Yang", "Hejin Wang", "Junwei Yang", "Zeren Zhang", "Ziwei Chen", "Huaibo Huang", "Jie Cao", "Ran He"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d641819d168eedc07270607da33f3197c15f2f90cf99f538c3839eed506f3308", "submitted_at": "2026-09-26T14:35:45Z", "updated_at": "2026-09-26T14:35:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32679v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.32679", "identifiers": {"arxiv": "2609.32679"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32679"}], "name": "The GUI Is Not the State: Diagnosing State Aliasing in GUI World Models", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces StateAliasBench to isolate GUI world-model ambiguities in which matching visible screens and actions can lead to different futures. It augments frozen world models with structured state recovered from interaction history and evaluates the effect on prediction and agent execution.", "topics": ["world-models", "state-tracking", "evaluation"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32679"}
{"aliases": [], "authors": ["Xin Yan", "Zhengbo Jiao", "Jiaqi Liu", "Zhenglin Wan", "SiYuan Ma", "Xuliang Yu", "Tianyi Jiang", "Chubin Zhang", "Pengfei Zhou", "Wangbo Zhao", "Xingrui Yu", "Bo An", "Yang You", "Ivor Tsang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4083bbe61c846d93663cd41d6335a2d6a54e361df9c3bdcd636b312c436d1a4", "submitted_at": "2026-09-26T16:16:08Z", "updated_at": "2026-09-26T16:16:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32750v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.32750", "identifiers": {"arxiv": "2609.32750"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32750"}], "name": "CUA-Sandbox: Efficient Environments for Computer-Use Agent Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CUA-Sandbox, which separates each rollout’s mutable state from shared initialized software runtimes. State-scoped execution and transactional resets or branches aim to reduce training-environment costs while retaining original interfaces and evaluators.", "topics": ["sandbox", "reinforcement-learning", "efficiency"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32750"}
{"aliases": [], "authors": ["Jerry Huang", "Sarvesh Babu", "Matt Van Buren", "Alexander Wang", "Pranav Pillai", "Arush Jain", "James P. Burton", "Julia Hockenmaier"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c3b0ca282c508618f08489175452c33ad681e0cb32af930681aa89ce795106af", "submitted_at": "2026-09-26T18:07:58Z", "updated_at": "2026-09-26T18:07:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32835v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.32835", "identifiers": {"arxiv": "2609.32835"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.32835"}], "name": "FinancialAuditBench: Benchmark Construction under Differential Privacy Using Real-World Priors", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-26", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32835"}
{"aliases": [], "authors": ["Asif Shahriar", "Md Nafiu Rahman", "Sadif Ahmed", "Farig Sadeque", "Md Rizwan Parvez"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "802342ce2c486c30c84c33c897becb0348831034d4d11f8d3d3861ec0380b65c", "submitted_at": "2026-09-26T20:09:16Z", "updated_at": "2026-09-26T20:09:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32915v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.32915", "identifiers": {"arxiv": "2609.32915"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32915"}], "name": "AgentTell: Behavioural Side-Channel Leakage in Browser-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-26", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AgentTell, a benchmark for privacy leakage through browser actions after an agent learns a private fact on another website. It examines whether agents choose revealing options despite nondisclosure instructions and whether their final reports accurately describe that behavior.", "topics": ["safety-security", "evaluation", "memory"], "updated_at": "2026-09-26", "url": "https://arxiv.org/abs/2609.32915"}
{"aliases": [], "authors": ["Ben Hagag", "William L. Anderson", "Srija Chakraborty", "Christian Schroeder de Witt"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c380079220050fbe354509238a5a79c7df2917a245490e4ec0cff512f7565ab", "submitted_at": "2026-09-27T02:20:08Z", "updated_at": "2026-09-27T02:20:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.33102v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T12:58:20.217517Z", "role": "primary", "sha256": "f275c47a8b6d349a70e39a9e55eea5a09d99d5f35cc896100f3bcfd5b0923e0b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202609200000+TO+202610201255%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:28.833839+00:00", "role": "primary", "sha256": "3ea667f4187ba601011e5ff00c4d63676b6d9903d12b80a0d7d5dd9920f00c65", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2609.33102v1"}], "id": "arxiv:2609.33102", "identifiers": {"arxiv": "2609.33102"}, "kind": "paper", "last_verified_at": "2026-10-20T13:12:28.833839+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.33102"}], "name": "ORBIT: A Framework for Multi-Agent Safety and Security Evaluations", "notes": "BrowserART uses an accessibility-tree action interface, while the desktop adapter uses mouse, keyboard and screenshots. The adapted tasks are not claimed to preserve the constructs of their source benchmarks. Experiments cover a representative configuration slice with fixed attack templates; the OS-Harm results are imported from prior work using a different model and stack. Browser defense tables also differ in baseline versus defended judges and in judged benign denominators. Human judge validation used one grader. These results do not establish a universally effective defense. The linked repository and its code/data licenses were not independently inspected, and no experiment was run.", "published_at": "2026-09-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Inspect-based framework that varies multi-agent roles, communication, attacks and defenses, with BrowserGym and virtual-desktop GUI adapters alongside coding and other tool scenarios.", "topics": ["evaluation", "safety-security", "web-navigation"], "updated_at": "2026-09-27", "url": "https://arxiv.org/abs/2609.33102"}
{"aliases": [], "authors": ["Fuyao Zhang", "Xuan Wang", "Zherui Li", "Jiaming Zhang", "Longtao Huang", "Wei Yang Bryan Lim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "563661718e6e64d62f0601bbc5c214828219b9495a729ae685ba1ae9599332f3", "submitted_at": "2026-09-27T07:15:57Z", "updated_at": "2026-09-27T07:15:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.33304v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.33304", "identifiers": {"arxiv": "2609.33304"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.33304"}], "name": "Relevance Does Not Imply Applicability: Experience Activation for Personal GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ExpActivator, which matches the current GUI state to relevant historical states before supplying past actions as guidance. A separate activation rule decides whether a recurring user intent is warranted by the current time and situation or should be withheld.", "topics": ["memory", "planning", "training-free"], "updated_at": "2026-09-27", "url": "https://arxiv.org/abs/2609.33304"}
{"aliases": [], "authors": ["Keliang Li", "Heng Wang", "Chen Hu", "Daxin Jiang", "Hong Chang", "Shiguang Shan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88daacf5fda62f59a08b504a633ad984ee1d03791bba24e89c223c0686adf34e", "submitted_at": "2026-09-27T15:11:30Z", "updated_at": "2026-09-27T15:11:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.33646v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "id": "arxiv:2609.33646", "identifiers": {"arxiv": "2609.33646"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.33646"}], "name": "Probe to Act: Elevating Browser-Use Agent via Active Visual Probing", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-27", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Probe to Act, which lets browser agents request visual evidence linking DOM handles and screen regions before committing an operation. It retains selected interaction evidence as memory and studies both prompted use and distillation into open models.", "topics": ["grounding", "memory", "web-navigation"], "updated_at": "2026-09-27", "url": "https://arxiv.org/abs/2609.33646"}
{"aliases": [], "authors": ["Shaoqing Zhang", "Kehai Chen", "Xuefeng Bai", "Zhuosheng Zhang", "Pengfei Zhang", "Yang Xiang", "Min Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f4116455785579ece9a2ff1bae304b62f1cbe13a91db8d146bd281a27d19eb0", "submitted_at": "2026-09-28T01:48:48Z", "updated_at": "2026-09-29T09:05:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.34113v2"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.34113", "identifiers": {"arxiv": "2609.34113"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.34113"}], "name": "GUITAR: Structured Failure Diagnosis of GUI Agents via State Transitions", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GUITAR, which groups visually different screens into functional states and analyzes agent failures over a state-transition graph. It uses the resulting bottlenecks to diagnose recurring mobile and web errors and evaluate targeted guidance.", "topics": ["evaluation", "observability", "robustness"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.34113"}
{"aliases": [], "authors": ["Tien Tran", "Namho Koh", "Daiki E. Matsunaga", "Ayush Jain", "Kee Eung Kim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ee7ddc170afcd2fa09d1c8d12a07bbd6aac692bc12ffc27bd39f1de0c912dfa", "submitted_at": "2026-09-28T02:21:13Z", "updated_at": "2026-09-28T02:21:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34139v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.34139", "identifiers": {"arxiv": "2609.34139"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.34139"}], "name": "Same Tasks, Different Apps: Why Mobile GUI Agents Fail to Generalize?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AnyAppBench, a live Android benchmark that holds user goals fixed while changing the application used to achieve them. It combines category-controlled task pairs with failure annotations to distinguish task understanding from application-specific execution skill.", "topics": ["evaluation", "robustness", "mobile-navigation"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.34139"}
{"aliases": [], "authors": ["Ziqiang Wang", "Li Gu", "Zhixiang Chi", "Linlian Jiang", "Zihuan Jiang", "Linqiang Guo", "Siobhan Reid", "Zhi Liu", "Yang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fbaeb0b4f891d64cc12a3d5663b6804df6c7db82e6a7ccfd54aa10e837010de", "submitted_at": "2026-09-28T05:00:45Z", "updated_at": "2026-09-28T05:00:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34321v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.34321", "identifiers": {"arxiv": "2609.34321"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.34321"}], "name": "One Rollout Is All You Get: Fully Test-Time Adaptation for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Defines a test-time adaptation setting with one counted attempt per arriving GUI task and no ground-truth feedback. Its SOLO method filters or relabels completed episodes, then updates a small adapter from a short window of the agent’s own experience.", "topics": ["memory", "evaluation", "robustness"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.34321"}
{"aliases": [], "authors": ["Chanhee Park", "Jeongho Yoon", "Sungbin Han", "Hyeonseok Moon", "Heuiseok Lim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "beabbfc9e6f81da198650fd25ed0501309d0812b55a3fdc17ff480070c03a995", "submitted_at": "2026-09-28T06:45:01Z", "updated_at": "2026-09-28T06:45:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34428v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.34428", "identifiers": {"arxiv": "2609.34428"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.34428"}], "name": "AgentHop: A Diagnostic Benchmark for Agentic Multi-Hop Scientific Question Answering", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.34428"}
{"aliases": [], "authors": ["Bingo Zhang", "Haochuan Lu", "Zongjie Li", "Genjian Li", "Ari Yu Zhang", "Chaozheng Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f81fd07262dccf6783729e4336fbd77835f2fcdea030eef8dc2272e7ef82b8fc", "submitted_at": "2026-09-28T09:49:25Z", "updated_at": "2026-09-28T09:49:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34769v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.34769", "identifiers": {"arxiv": "2609.34769"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.34769"}], "name": "LongPuzzleBench: Evaluating GUI Agents on Long-Horizon Visual Puzzles", "notes": "The annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced. Agents execute real GUI moves in six puzzle games; the tasks diagnose computer-control planning but remain game-based rather than general productivity workflows.", "published_at": "2026-09-28", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "LongPuzzleBench evaluates coherent long-horizon planning on persistent visual puzzle boards through native GUI actions. It compares GUI-only and code-assisted agents and uses controlled hints and memory diagnostics to examine decisions that preserve or eliminate future solutions.", "topics": ["evaluation", "planning", "long-horizon", "state-tracking"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.34769"}
{"aliases": [], "authors": ["Ruozhao Yang", "Mingfei Cheng", "Xiaofei Xie"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9bc3b57957e6c4f1af399379c4275366563b94a8b8fe01ec83e54715cfc8d11", "submitted_at": "2026-09-28T11:53:24Z", "updated_at": "2026-09-28T11:53:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34974v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.34974", "identifiers": {"arxiv": "2609.34974"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.34974"}], "name": "Before Acting, Change the State: Prospective State Intervention for Web Agents under Deceptive Interfaces", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Veer, a runtime defense that changes task-relevant web state when an otherwise valid proposed action would produce an unauthorized consequence. It plans a prospective intervention and grounds and verifies its execution before the base agent continues.", "topics": ["safety-security", "planning", "robustness"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.34974"}
{"aliases": [], "authors": ["Qiankun Li", "Yuechen Zhang", "Bowen Chen", "Shilinlu Yan", "Zhenhong Zhou", "Kun Wang", "Li Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "48cd1fc30abc6338f03d802397ab9ca1cc580bcb7b131a85e15cf23b92fd95c1", "submitted_at": "2026-09-28T12:06:18Z", "updated_at": "2026-09-28T12:06:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35002v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.35002", "identifiers": {"arxiv": "2609.35002"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.35002"}], "name": "Still There, No Longer Seen: Exposing Compression-Induced Risk in Large Vision-Language Models", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.35002"}
{"aliases": [], "authors": ["Anton Emelyanov", "Maria Tikhonova", "Zaven Martirosian", "Sergei Averkiev", "Alena Fenogenova"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5090c6b7879e38102b1180ee1e4dfd1552af777172e2f94bc0f9a684fb59dc7", "submitted_at": "2026-09-28T12:24:23Z", "updated_at": "2026-09-28T12:24:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35026v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.35026", "identifiers": {"arxiv": "2609.35026"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.35026"}], "name": "WebPageBench: Event-Level Verification and Controlled UI-Variant Generation for Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Verifies web tasks through typed interface events and generates controlled UI variations while preserving instructions and success conditions.", "topics": ["evaluation", "robustness"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.35026"}
{"aliases": [], "authors": ["Rui Zhong", "Yu Li", "Zheyu Yan", "Cheng Zhuo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49fefb832ad64774d759bfbf4a9419152f6aad305bef21683fca1dfbe53f5307", "submitted_at": "2026-09-28T14:10:25Z", "updated_at": "2026-09-29T14:19:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.35232v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.35232", "identifiers": {"arxiv": "2609.35232"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.35232"}], "name": "Beyond Selection: Token Parameterization for Extreme Visual Token Compression", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.35232"}
{"aliases": [], "authors": ["Xiao Zhang", "Wang Zeng", "Sheng Jin", "Wentao Liu", "Chen Qian", "Shichao Kan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b8fbdc2d5ddbd198e24efbeada59293cf0c0e86f38da39fe7c9de99b6621e334", "submitted_at": "2026-09-28T15:21:37Z", "updated_at": "2026-09-28T15:21:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35394v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.35394", "identifiers": {"arxiv": "2609.35394"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.35394"}], "name": "Rethinking Visual Token Compression for Video Large Language Models: A Simple Yet Strong Baseline", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.35394"}
{"aliases": [], "authors": ["Yangqin Jiang", "Lingrui Xu", "Chao Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ffb5951994cbccef0700ce850a1fae096a59b87334070750caf0b7b42d61c4b", "submitted_at": "2026-09-28T17:26:56Z", "updated_at": "2026-10-05T01:08:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.35671v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2609.35671", "identifiers": {"arxiv": "2609.35671"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.35671"}], "name": "PhoneCLI: From App Interfaces to Callable Commands for Mobile Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PhoneCLI, which explores mobile app interfaces and compiles navigation paths into reusable commands without internal app APIs or instrumentation. At execution time it verifies compiled paths and falls back to a visual agent for open-ended interaction or failed navigation.", "topics": ["automation", "efficiency", "mobile-navigation"], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2609.35671"}
{"aliases": [], "authors": ["Zhan Qu", "Hui Zang", "Ran Chen", "Tao Wang", "Shengyu Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "903b60e1ffb8a7b695f034d8ca1cc827e7d601b506d39a1cbbde404921f364b9", "submitted_at": "2026-07-31T02:08:54Z", "updated_at": "2026-07-31T02:08:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35775v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.35775", "identifiers": {"arxiv": "2609.35775"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.35775"}], "name": "SPECTRA: On-Device Cognitive Perturbation and Trajectory Analysis for Autonomous Edge-Cloud GUI Grounding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-07-31", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-07-31", "url": "https://arxiv.org/abs/2609.35775"}
{"aliases": [], "authors": ["Xunjian Yin", "Tianchen Guan", "Jinao Wang", "Weili Cao", "Daisy Xinlei Lin", "Royce Cheng-Yue", "Keagan Long", "Kyle Wong", "Bhuwan Dhingra", "Xiangjun Wang", "Shuyan Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "197cfb8424584955bc85602d6559037ad31ee815f89b45dc8f9f18c8636552ad", "submitted_at": "2026-09-20T16:16:23Z", "updated_at": "2026-09-20T16:16:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35814v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2609.35814", "identifiers": {"arxiv": "2609.35814"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.35814"}], "name": "Constructing Challenging Browser-Use Tasks by Controlled Environment Interventions", "notes": "The annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.", "published_at": "2026-09-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "BreakingWeb pairs solvable browser tasks with controlled environment interventions that preserve the goal, target and backend success criterion. The paired evaluations isolate robustness failures, including agents declaring completion without the required state change.", "topics": ["evaluation", "robustness", "data-generation", "state-tracking"], "updated_at": "2026-09-20", "url": "https://arxiv.org/abs/2609.35814"}
{"aliases": [], "authors": ["Ziqi Zhang", "Shaohui Li", "Bing Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7efd2eafa1984d7c8f4bf6ce79a94cd2ca4189da330bdd056b4be9c07f7f9dc1", "submitted_at": "2026-09-28T04:02:07Z", "updated_at": "2026-09-28T04:02:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35904v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2609.35904", "identifiers": {"arxiv": "2609.35904"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.35904"}], "name": "Structured Interaction, Visual Localization, and Robust Execution for Complex Web Tasks: A Technical Report on the WebRetriever Challenge", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes a WebRetriever agent that combines semantic webpage operations with visual localization when structured representations are insufficient. Hierarchical context management and fault-aware execution support processing tasks across concurrent browsers.", "topics": ["web-navigation", "grounding", "robustness", "memory"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.35904"}
{"aliases": [], "authors": ["Meituan LongCat Team", "He Zhu", "Yue Xu", "Wanli Wu", "Haolin Ren", "Yuxin Bian", "Jiarui Zhao", "Rongzhi Zhang", "Quanchi Weng", "Jinghao Cui", "Yu Fan", "Yuhan Liu", "Yunhu Ye", "Jiyuan Ren", "Fengcheng Yuan", "Zhao Yang", "Jiacheng Zhang", "Yuchuan Dai", "Ruixuan Xiao", "Haozhe Sun", "Xiangyuan Liu", "Cheng Sun", "Yao Du", "Yiming Hao", "Hongbo Guo", "Shuo He", "Lei Wang", "Xunliang Cai", "Yan Chen", "Fan Yang", "Lingchuan Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b90b3f3393dd2fe8ac495fa26f696fda07022bc179a9e810d678af50e06303ea", "submitted_at": "2026-09-28T18:22:38Z", "updated_at": "2026-09-28T18:22:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36071v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.36071", "identifiers": {"arxiv": "2609.36071"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.36071"}], "name": "LongCat-DeepResearch Technical Report", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.36071"}
{"aliases": [], "authors": ["Jianxing Chen", "Xiao Yu", "Shipra Agrawal", "Zhou Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e92c63d4cf58375aa7068570f8a98c145e24f8d393c5807c0aff3ed67f18b9f0", "submitted_at": "2026-09-28T20:01:35Z", "updated_at": "2026-09-28T20:01:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36201v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.36201", "identifiers": {"arxiv": "2609.36201"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36201"}], "name": "SCOUT: Synergizing Reasoning and Tool-Use for Computer-Use Safety", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-28", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SCOUT, a safety verifier that first derives task-specific completion and safety rubrics from a computer-use trajectory, then actively probes the resulting environment for evidence. It evaluates whether these two stages detect harmful consequences that screenshots alone can miss.", "topics": ["safety-security", "evaluation", "tool-integration"], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.36201"}
{"aliases": [], "authors": ["Ziyi Yin", "Sangmin Woo", "Kang Zhou", "Sungyeon Kim", "Aosong Feng", "Haibo Ding", "Jun Huan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd3f4c71b55b83419395f45e04deb6b33f3adbe27403f29127e84400a26a26d7", "submitted_at": "2026-09-28T22:34:43Z", "updated_at": "2026-09-28T22:34:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36352v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.36352", "identifiers": {"arxiv": "2609.36352"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.36352"}], "name": "StructRL: Online Structured Reinforcement Learning for Long-Horizon Vision-Language-Action Tasks", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-28", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-28", "url": "https://arxiv.org/abs/2609.36352"}
{"aliases": [], "authors": ["Dongchan Shin", "Xing Han Lù", "Jiaqi Deng", "Jay Gala", "Tomás Vergara Browne", "Jaewon Moon", "Fengyuan Liu", "Alexandre Drouin", "Siva Reddy", "Alexandre Lacoste"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2fe39365133228daa770445f19002f6abe34ca1151cf4952a225ba0001de5250", "submitted_at": "2026-09-29T01:46:48Z", "updated_at": "2026-09-29T01:46:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36488v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.36488", "identifiers": {"arxiv": "2609.36488"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36488"}], "name": "AdaptArena: Evaluating Test-Time Personalization of Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AdaptArena to test whether web agents can infer unstated user preferences from relevant past trajectories and carry those preferences into new tasks. Its AdaptiveAgent baseline separates preference-retrieval challenges from failures in downstream web execution.", "topics": ["evaluation", "memory", "web-navigation"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.36488"}
{"aliases": [], "authors": ["Xiaokang Ye", "Siddhant Hitesh Mantri", "Zimeng Chen", "Edward Zhang", "Zhaoxu Zheng", "Yuanheng Li", "Yizhao Chen", "Tianyang Huang", "Lianhui Qin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "56f5f6324640b20830f7ab36124d0f29a7b991c1b7eec27ef4bb3bbe90b6b6a4", "submitted_at": "2026-09-29T05:49:37Z", "updated_at": "2026-09-29T05:49:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36777v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.36777", "identifiers": {"arxiv": "2609.36777"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.36777"}], "name": "Code4Scene: Benchmarking Coding Agents for Constructing and Editing 3D Scenes", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.36777"}
{"aliases": [], "authors": ["Bin Kang", "Jiarui Ouyang", "Li Jiang", "Bin Chen", "Zhuotao Tian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3af30198735ca9b9d192e15acbb7f8a37fe222fa2c0c1d2ccff448ef5e79f55", "submitted_at": "2026-09-29T07:41:34Z", "updated_at": "2026-09-29T07:41:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36923v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.36923", "identifiers": {"arxiv": "2609.36923"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36923"}], "name": "PrecogUI: Proactive GUI Agents via Pre-cognitive Simulation and Experience Retrieval", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents PrecogUI, which retrieves prior anomaly and success patterns and predicts symbolic UI transitions before choosing actions. It also introduces InterfereBench, generated for evaluating long GUI workflows with disruptive events and recovery requirements.", "topics": ["memory", "world-models", "robustness", "long-horizon"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.36923"}
{"aliases": [], "authors": ["Hyewon Suh", "Thanh Minh Nguyen", "Chih-Lun Lee", "Darrow Hartman", "Lizhao Liu", "Xin Eric Wang", "Ang Li", "Jiachen Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fdd9890482c07ef44f6e274c802bae5decd9d883b381545b3034ebd4001cdb34", "submitted_at": "2026-09-29T07:44:22Z", "updated_at": "2026-09-29T07:44:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36927v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.36927", "identifiers": {"arxiv": "2609.36927"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36927"}], "name": "Neuro-Symbolic Computer Use: Learning Reusable Policies for Reliable and Efficient Execution", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Learns reusable computer-use policies that encode stable workflow logic in programs while delegating grounding and state checks to neural models. Policy iteration executes variants, diagnoses failures, revises the program, and adds verification before state-changing actions.", "topics": ["automation", "workflow", "efficiency"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.36927"}
{"aliases": [], "authors": ["Shengtian Yang", "Ziyu Xiong", "Kaibing Yang", "Guangfeng Cai", "Yewen Li", "Peng Jiang", "Gai Kun", "Qingpeng Cai", "Lei Feng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d45cee8117e302aa190d38c8298bf479936a5a27aa9b85716a76a2a90499069c", "submitted_at": "2026-09-29T07:51:57Z", "updated_at": "2026-09-29T07:51:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36939v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.36939", "identifiers": {"arxiv": "2609.36939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36939"}], "name": "SCA: Spatial Credit Assignment for Reinforcement Learning of GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Spatial Credit Assignment, which uses click coordinates to refine group-relative reinforcement-learning updates for GUI grounding. It distinguishes failed clicks by spatial evidence while leaving the deployed action policy unchanged.", "topics": ["grounding", "reinforcement-learning", "evaluation"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.36939"}
{"aliases": [], "authors": ["Shinan Zhang", "Tao Zhang", "Qihui Zhu", "Mengjie Zhang", "Dong Jin", "Yunpeng Hou", "Shuangwu Chen", "Xiaobin Tan", "Quan Zheng", "Jian Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67dbd96dee815b25ef15fa0c83105008bbdb6f60a22c1927244e5b3b019efb98", "submitted_at": "2026-09-29T08:36:55Z", "updated_at": "2026-09-29T08:36:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.37017v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.37017", "identifiers": {"arxiv": "2609.37017"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.37017"}], "name": "LatCom: Cross-Agent Latent Compression for Efficient Multi-Agent Collaboration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.37017"}
{"aliases": [], "authors": ["Mei Wu", "Rui Xie", "Runyu Zhang", "Yuqiang Li", "Tianfan Fu", "Bo Chen", "Kai Yu", "Xin Chen", "Lu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e032f67257a52dbc59310d921d802e4dc85af3af110d22bb6d005d202bbaa6b2", "submitted_at": "2026-09-29T08:58:51Z", "updated_at": "2026-09-29T08:58:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.37053v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2609.37053", "identifiers": {"arxiv": "2609.37053"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.37053"}], "name": "MatToolBench: Benchmarking Multimodal Agents in Real-World Materials Science Workflows", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MatToolBench, a Windows-based materials-science benchmark combining GUI operation, scientific-software scripting, and database queries. Expert-defined subcriteria support partial-credit scoring, while the evaluation examines domain knowledge gaps and cross-tool handoff failures.", "topics": ["science", "evaluation", "cross-app"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.37053"}
{"aliases": [], "authors": ["Giulio Segalini", "Zhi Wen Soi", "Jérémie Decouchant", "Lydia Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "945184e8fc790979a1f7e466ef6b4e727ddd11c3f285be0371fa4dda42615f40", "submitted_at": "2026-09-29T10:01:52Z", "updated_at": "2026-09-29T10:01:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.37176v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.37176", "identifiers": {"arxiv": "2609.37176"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.37176"}], "name": "Absorbed in Inertia: Activation Analysis for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies repetitive, unproductive actions in desktop agents through their internal activation states. It proposes Reset, Reroute, Restore, a context intervention that temporarily resets an agent’s trajectory and later restores it to help escape repeated ineffective behavior.", "topics": ["robustness", "observability", "memory"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.37176"}
{"aliases": [], "authors": ["Hongcheng Gao", "Hailong Qu", "Yu Lei", "Henghui Sun", "Haoyang Li", "Yipeng Wei", "Naihao Xue", "Xiaohan Yu", "Zhuo Tao", "Yihe Zang", "Yajiao Wang", "Jingyi Tang", "Yi Li", "Jingjing Zhou", "Jie Luo", "Bohan Zeng", "Chengyu Shen", "Hao Jiang", "Chong Chen", "Bowen Qu", "Olive Huang", "Zeqiang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aab7f586cb58a1f2d20267fd810ba34d6d4d166e3da715d2b3ff0b77e464473e", "submitted_at": "2026-09-29T14:32:35Z", "updated_at": "2026-10-09T08:15:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.37686v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2609.37686", "identifiers": {"arxiv": "2609.37686"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.37686"}], "name": "EngiWorld: What Can Frontier Agents Deliver in Professional Engineering Environments?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces professional engineering tasks across software domains, with GUI and CLI access and artifact checks for geometric, physical, and design validity.", "topics": ["evaluation", "science", "cross-app"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2609.37686"}
{"aliases": [], "authors": ["Tongbo Chen", "Junbo Niu", "Zhengxi Lu", "Niu Lian", "Fei Tang", "Yuchen Yan", "Yike Hong", "Yong Du", "Yizhou Liu", "Bofan Chen", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b394677ebe37782950cfc56c0be65ca7a93f34e5550ec9425d91921bf1799847", "submitted_at": "2026-09-29T16:57:25Z", "updated_at": "2026-09-29T16:57:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38008v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.38008", "identifiers": {"arxiv": "2609.38008"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.38008"}], "name": "HybridCUA: Learning to Orchestrate GUI and CLI for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains computer agents on verified GUI, command-line, and interleaved trajectories to learn when and how to combine interaction modes.", "topics": ["tool-integration", "reinforcement-learning", "data-generation"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.38008"}
{"aliases": [], "authors": ["Yongjiang Wu", "Junyuan Zhang", "Ada Chen", "Kuiyi Gao", "Wenxuan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f14844abefab6b4eb464a1c7675f9de7f460849872ebb776230724ed5828feed", "submitted_at": "2026-08-07T07:51:13Z", "updated_at": "2026-08-07T07:51:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38184v1"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.38184", "identifiers": {"arxiv": "2609.38184"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.38184"}], "name": "RealGUINoise: An Interactive Cross-Platform Benchmark for GUI Agent Robustness under Real-World Interface Noise", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-08-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates interactive GUI agents under varied realistic interface noise, comparing task completion, safety, and behavior with clean-environment runs.", "topics": ["robustness", "evaluation", "safety-security"], "updated_at": "2026-08-07", "url": "https://arxiv.org/abs/2609.38184"}
{"aliases": [], "authors": ["Yunan Lu", "Shuang Xie", "Meghna Allamudi", "Mingyu Zhao", "Han Li", "Lingyun Wang", "Zhou Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f02e78f50d1c6065dfd8ebca78f2cc83421b67bc6cc74616c2861d42370e0cb2", "submitted_at": "2026-09-29T18:53:23Z", "updated_at": "2026-09-29T18:53:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38397v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2609.38397", "identifiers": {"arxiv": "2609.38397"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.38397"}], "name": "SimTrace: Grounded Multimodal User Trajectories Generation for Online User Modeling", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2609.38397"}
{"aliases": [], "authors": ["Hongbo Zhang", "Liuyang Song", "Quanquan Li", "Daqian Yang", "Yan Wen", "Zhengtao Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb910e64ebb65bd49e8729e4373c1c81ce4b2d0a32647a5abaf15f52737e578d", "submitted_at": "2026-09-30T02:11:02Z", "updated_at": "2026-09-30T02:11:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38778v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.38778", "identifiers": {"arxiv": "2609.38778"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.38778"}], "name": "Action Conditioned Bisimulation For GUI Agent Memory", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Defines a GUI-memory merge rule that treats two states as equivalent only when shared actions lead to matching outcomes and successor states. The training-free method replaces observation-similarity merging and is tested in closed-loop MiniWoB++ interaction.", "topics": ["memory", "web-navigation"], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2609.38778"}
{"aliases": [], "authors": ["Terumi Chiba", "Guangzhi Sun", "Zheqi Yuan", "Chao Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bded059a37101b0e6a91ba662e0d70e5e010ddb8c3557598dff2ba82afa45e34", "submitted_at": "2026-09-30T03:19:25Z", "updated_at": "2026-09-30T03:19:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38867v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2609.38867", "identifiers": {"arxiv": "2609.38867"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.38867"}], "name": "Talk2Agent: Benchmarking Voice Interfaces for Text Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Talk2Agent, a benchmark of spoken instructions for computer-use tasks that measures whether voice interfaces preserve task-critical information. It pairs human speech with an execution-free intention-based metric and checks how that metric relates to downstream task completion.", "topics": ["evaluation", "robustness"], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2609.38867"}
{"aliases": [], "authors": ["Qisheng Su", "Hanchen Wang", "Guanru Zhu", "Huicheng Jiang", "Qiuyinzhe Zhang", "Kou Shi", "Zhen Fang", "Ziao Zhang", "Qingnan Ren", "Honglin Guo", "Zehui Chen", "Tao Gui", "Feng Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5374edb4b8b8014b7809db91cd0cadd50e7eb2b42e2fdc1c50890ba35370ca01", "submitted_at": "2026-09-30T04:07:17Z", "updated_at": "2026-10-04T14:53:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.38923v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.38923", "identifiers": {"arxiv": "2609.38923"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.38923"}], "name": "GraphForge: Training Working Agents with Graph-Anchored Workspace Synthesis", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-04", "url": "https://arxiv.org/abs/2609.38923"}
{"aliases": [], "authors": ["Xu Huang", "Ye Huang", "Zijun Liao", "Yuwei Niu", "Xiaojie Li", "Menghan Zhou", "De Wen Soh", "Xiaotong Li", "Daquan Zhou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b09fcb3bd94d9b1106490e950679b24b28a78802fa4342891b937b00e7e60cdb", "submitted_at": "2026-09-30T08:00:20Z", "updated_at": "2026-09-30T08:00:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.39222v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "id": "arxiv:2609.39222", "identifiers": {"arxiv": "2609.39222"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.39222"}], "name": "DC-SAE: Deep Compression Semantic Autoencoder for Faster Diffusion Convergence", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2609.39222"}
{"aliases": [], "authors": ["Shengjie Jin", "Hengbo Xu", "Zelong Sun", "YuJie Guo", "Zhiwu Lu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b7206bc9c4842e6488955151a161a526dbbf81407ccb83865029cfef02542fe9", "submitted_at": "2026-09-30T08:49:03Z", "updated_at": "2026-09-30T08:49:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.39306v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2609.39306", "identifiers": {"arxiv": "2609.39306"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.39306"}], "name": "ReSAIL: Mitigating Collapse in Iterative Agent Self-Distillation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2609.39306"}
{"aliases": [], "authors": ["Bo Han", "Qianyi Wang", "Shuai Liu", "Xiong Zifan", "Changqiao Wu", "Yuanfa Li", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Heng Qu", "Yunpeng Song", "Zhongmin Cai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3594620e97b41f3a7481871588b63ffa4aa8822e8beb42334b2cd888bcfcfaaf", "submitted_at": "2026-09-30T11:45:51Z", "updated_at": "2026-09-30T11:45:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.39547v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.39547", "identifiers": {"arxiv": "2609.39547"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.39547"}], "name": "Learning Reliable GUI Agents under Imperfect Priors", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Combines structured GUI exploration with training designed to tolerate imperfect retrieved experience. It builds state-transition graphs and synthetic task trajectories, then injects realistic forms of interface drift so agents learn when to rely on or reject prior guidance.", "topics": ["data-generation", "memory", "robustness"], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2609.39547"}
{"aliases": [], "authors": ["Dingyuan Dai", "Heli Qi", "Lei Liu", "Yinxi Li", "Baiding Chen", "Zijun Dou", "Qingcheng Zeng", "Qi Kang", "Oliver Sun", "Eric Wang", "Bo Zhou", "Haixin Wang", "Yufan Du", "Shi Bo", "Ruihan Lin", "Mengqi Yuan", "Dunjie Lu", "Steven Dillmann", "Yiming Shi", "Tina Su", "Amy Xin", "Minghao Liu", "Xi Wang", "Xu Huang", "Ge Zhang", "Pengyu Nie", "Zhen Yang", "Jie Tang", "Juanzi Li", "Weihao Xuan", "Tianyu Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "690cd75934bae65311240cd4e6db7b8fd607056bd60799bf70cddd03efda9261", "submitted_at": "2026-09-30T15:03:05Z", "updated_at": "2026-09-30T15:03:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.39903v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.39903", "identifiers": {"arxiv": "2609.39903"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.39903"}], "name": "OSWorld-Science: A Benchmark of Computer Use Agents for Learning and Using Scientific Software", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Benchmarks computer agents on scientific-software workflows with task-specific checks of application states and generated research artifacts.", "topics": ["evaluation", "science"], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2609.39903"}
{"aliases": [], "authors": ["Yulong Liu", "Xiaotian Han", "Junyuan Shang", "Yuchen Ding", "Zhenyu Zhang", "Shuohuan Wang", "Guibo Zhu", "Sirui Han", "Dianhai Yu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "46a466b935dea473ae6aa13fa1c9c24fd74e40c6169616fedd713b2f05dd3e88", "submitted_at": "2026-09-30T15:12:10Z", "updated_at": "2026-09-30T15:12:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.39924v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2609.39924", "identifiers": {"arxiv": "2609.39924"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2609.39924"}], "name": "CoVisco: Codec-Native Vision Encoder with Native Token Compression for Unified Image-Video Understanding", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2609.39924"}
{"aliases": [], "authors": ["Yong Du", "Tongbo Chen", "Zhengxi Lu", "Yizhou Liu", "Bofan Chen", "Tao Jiang", "Wenhao Xu", "Yongliang Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8f3745b6b497c8819dc990bdb5d74e7af463d23654cbc7cdd76ade0faf2b3e0c", "submitted_at": "2026-09-30T17:40:09Z", "updated_at": "2026-10-01T17:50:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.40253v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.40253", "identifiers": {"arxiv": "2609.40253"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.40253"}], "name": "ComputerSD: Online Self-Distillation from Real-Time Feedback for Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ComputerSD, which turns feedback from executed GUI transitions into online self-distillation guidance and step-level scores. A GUI analyzer supplies these signals, which are combined with trajectory-level reinforcement learning in asynchronous training.", "topics": ["reinforcement-learning", "evaluation"], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2609.40253"}
{"aliases": [], "authors": ["Pranjal Aggarwal", "Lawrence Keunho Jang", "Sean Welleck", "Daniel Fried", "Ruslan Salakhutdinov", "Jing Yu Koh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "502912cdc35afd61beeb0202d55fc8feefcd92cd3719cd0e85bc71561d6f5a3c", "submitted_at": "2026-09-30T17:48:06Z", "updated_at": "2026-09-30T17:48:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.40284v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2609.40284", "identifiers": {"arxiv": "2609.40284"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.40284"}], "name": "cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-09-30", "relations": [], "review_status": "curated", "scope": "core", "summary": "Standardizes virtual machines, interaction pipelines, and agent interfaces to compare computer-use speed, cost, and task performance.", "topics": ["evaluation", "efficiency"], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2609.40284"}
{"aliases": [], "authors": ["Gene Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bc7f3c818eb8b0813f3855e0f006aa63ce2fcc8f9ab6bbb6e65c14bf485aa444", "submitted_at": "2026-09-03T03:40:18Z", "updated_at": "2026-09-03T03:40:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.00042v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.00042", "identifiers": {"arxiv": "2610.00042"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.00042"}], "name": "Zengram-Lite: An In-Browser Agentic-Memory Framework - Semantic Knowledge, Session Tracking, and Token-Budgeted Context", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-03", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-03", "url": "https://arxiv.org/abs/2610.00042"}
{"aliases": [], "authors": ["Amir Rafe", "Subasish Das"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb4850824fa40587169257ae1be12c1154c591a5daf692ec723c820addacd5ff", "submitted_at": "2026-09-29T19:57:59Z", "updated_at": "2026-10-06T22:03:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.00346v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.00346", "identifiers": {"arxiv": "2610.00346"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.00346"}], "name": "Benchmarking System One decision models against trained classifiers and language models for automated decision gates", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-29", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-06", "url": "https://arxiv.org/abs/2610.00346"}
{"aliases": [], "authors": ["Haokai Ma", "Chieh Lin", "Yupeng Qiu", "Ee-Chien Chang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "20dcdc313f032ac84dca7f7ba9e1b8af452fceb7774cd7ec6fea90fd893bbc7e", "submitted_at": "2026-09-30T17:59:02Z", "updated_at": "2026-09-30T17:59:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.00450v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.00450", "identifiers": {"arxiv": "2610.00450"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.00450"}], "name": "ZoneClaw: Mitigating Persistent Memory Attacks by Establishing Memory-Zoning in OpenClaw-Style Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2610.00450"}
{"aliases": [], "authors": ["Sahan Paliskara", "Nattaput Namchittai", "Andrew Lampinen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ab6947bcf11eb3b60e4c665646357374e351fdf3ce59311e53fe45689ad7900", "submitted_at": "2026-09-30T18:49:25Z", "updated_at": "2026-09-30T18:49:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.00583v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.00583", "identifiers": {"arxiv": "2610.00583"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.00583"}], "name": "Worse Together: How Performance Breaks Down in Multi-User Multi-Agent Teams", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-09-30", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-09-30", "url": "https://arxiv.org/abs/2610.00583"}
{"aliases": [], "authors": ["Geyi Yang", "Zikun Qu", "Xiang Li", "Zhiyong Wang", "Min Zhang", "Shipei Zeng", "Zhongxiang Dai"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c43224c865fba2a9f4c826a8e912ab0c07297716e4b7640bf312cce9c9c510e", "submitted_at": "2026-10-01T02:31:20Z", "updated_at": "2026-10-01T02:31:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.00948v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.00948", "identifiers": {"arxiv": "2610.00948"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.00948"}], "name": "GUI-HARVEST: Self-Improving GUI Agents through Evidence-Driven Harness Evolution", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Improves GUI-agent execution code by analyzing action effects and repeated-task failures, then validating bounded changes with frozen model weights.", "topics": ["automation", "observability", "robustness"], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2610.00948"}
{"aliases": [], "authors": ["Cheng Yang", "Yifan Wu", "Yutao Huang", "Zhaohua Zhang", "Beiduo Chen", "Muxi Chen", "Chenchen Zhao", "Hexuan Deng", "Haolin Yang", "Geyuan Zhu", "Sa Zhu", "Jianhuan Zhuo", "Qiuyong Xiao", "Jianhao Ruan", "Yiran Peng", "Jiayi Zhang", "Tian Ye", "Xinlei Yu", "Tianwen Jiang", "Jihong Zhang", "Yuyu Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecd40b78c6de41abb442705b993aad8ebf7b35741536a11e00ebbd60140d1f18", "submitted_at": "2026-10-01T07:22:53Z", "updated_at": "2026-10-01T07:22:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.01215v1"}}, "environments": ["desktop", "linux", "windows", "macos", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.01215", "identifiers": {"arxiv": "2610.01215"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.01215"}], "name": "AutoGUIWorld: Image Generators as Visual World Models for GUI Agent", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AutoGUIWorld, which combines task planning and image editing to synthesize action-grounded GUI trajectories without running the depicted software. It filters generated transitions and tests whether training on them improves agents operating real desktop and scientific applications.", "topics": ["world-models", "data-generation", "grounding"], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2610.01215"}
{"aliases": [], "authors": ["Chengguang Gan", "Zimeng He", "Yoshihiro Tsujii", "Ken-ichiro Kobayashi", "Hiroki Itoh", "Kotaro Funakoshi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9a5b97264b3928c8dc74b955cc4baa7005cd34b0ce9e42cbc48a63bc6dd1f5a7", "submitted_at": "2026-10-01T11:31:37Z", "updated_at": "2026-10-01T11:31:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.01491v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.01491", "identifiers": {"arxiv": "2610.01491"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.01491"}], "name": "Auditing Web Agent Evaluation on WebArena-Lite: Human Review of Outcomes and Trajectories", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Audits WebArena-Lite outcomes and trajectories with human review to identify automatic-evaluator false negatives and the first consequential execution errors. It also studies explicit execution-state memory and procedural guidance, showing why final task scores can miss substantial differences in agent behavior.", "topics": ["evaluation", "observability", "memory"], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2610.01491"}
{"aliases": [], "authors": ["Beining Wu", "Zihao Ding", "Jun Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e85a601c31c6d06831ed1c590deef29d1c365e5d4751a308e42e8496ef11b19", "submitted_at": "2026-10-01T14:34:32Z", "updated_at": "2026-10-01T14:34:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.01787v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.01787", "identifiers": {"arxiv": "2610.01787"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.01787"}], "name": "Not All Experience Belongs in the Weights: Component Routing for Self-Improving GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies how to route different parts of GUI experience into model weights or runtime context rather than treating whole trajectories as a single unit. It separates locators, procedures, state facts, and lessons, then predicts their destination from recurrence and dependence on the current state.", "topics": ["memory", "evaluation", "efficiency"], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2610.01787"}
{"aliases": [], "authors": ["A. Said Gurbuz", "Ahmed Nassar", "Sunghwan Hong", "Marc Pollefeys", "Peter W. J. Staar"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd66468c2670ec22216e1592719a05e0b56af69606237a50f2067a340172553e", "submitted_at": "2026-10-01T18:00:05Z", "updated_at": "2026-10-01T18:00:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02320v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.02320", "identifiers": {"arxiv": "2610.02320"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.02320"}], "name": "DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2026-10-01", "relations": [], "review_status": "curated", "scope": "core", "summary": "Generates dense desktop supervision by varying real application scenes and combining screenshots, accessibility trees, and window geometry.", "topics": ["data-generation", "grounding"], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2610.02320"}
{"aliases": [], "authors": ["Dhananjay Ashok", "Shantanu Agarwal", "Vivek Datla", "Jonathan May", "Alfy Samuel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "35ded32e7bac9a875a0fa41efad43fed5fb0aab54fade5cfc2a0628b1e666587", "submitted_at": "2026-10-01T22:30:08Z", "updated_at": "2026-10-01T22:30:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02542v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.02542v1"], "reason": "The retained version-specific arXiv HTML infobox explicitly labels this paper License: CC BY 4.0. This document declaration does not establish code, training-data or model-weight licenses."}}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:08:56Z", "role": "primary", "sha256": "d8c22c6b3f07d4156c38af9dd7ca7c06a5d0a11889a7350d821a020e21604590", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.02542v1"}], "id": "arxiv:2610.02542", "identifiers": {"arxiv": "2610.02542"}, "kind": "paper", "last_verified_at": "2026-10-20T15:08:56Z", "licenses": {"code": null, "data": null, "document": {"basis": "The retained version-specific arXiv HTML infobox explicitly labels this paper License: CC BY 4.0. This document declaration does not establish code, training-data or model-weight licenses.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T15:08:56Z", "url": "https://arxiv.org/html/2610.02542v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.02542"}], "name": "How To Train Your World Model: Fine-tuning vs RAG for LM-based World Modeling", "notes": "Adjacent cross-domain study with WebShop as its web evaluation. WebShop uses text-rendered search/product pages and an explicit valid-action menu, not screenshot control or live purchases. A frozen policy proposes five actions; a world model predicts their next observations and the policy scores them before one is executed in the simulator. Utility is normalized environment reward, while fidelity uses a model judge; neither is a universal GUI task-success rate. Fine-tuning and retrieval use collected training experience; the hybrid masks same-episode retrieval, which does not establish general contamination freedom. The retained Atom and v1 HTML abstract wording differs; original bibliographic metadata and its abstract digest are preserved.", "published_at": "2026-10-01", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Compares fine-tuned, retrieval-based and hybrid language-model world models for next-observation prediction and action selection in text environments including WebShop.", "topics": ["world-models", "planning", "evaluation", "web-navigation"], "updated_at": "2026-10-01", "url": "https://arxiv.org/abs/2610.02542"}
{"aliases": [], "authors": ["Yun-Yun Tsai", "Yuning Mao", "Shiqi Wang", "Junfeng Yang", "Sinong Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7ed117876ece99f55e241e6365d0e785b907f9a7a24cceb01b1fd757191a35c9", "submitted_at": "2026-10-02T00:17:11Z", "updated_at": "2026-10-02T00:17:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02617v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:31.520084+00:00", "role": "primary", "sha256": "f053c4c65f6443acbda26e25cf36f7b6e8ea45f2916535604229f7c4e62ed679", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:31.520084+00:00", "role": "primary", "sha256": "f053c4c65f6443acbda26e25cf36f7b6e8ea45f2916535604229f7c4e62ed679", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.02617", "identifiers": {"arxiv": "2610.02617"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:31.520084+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.02617"}], "name": "WebUIProof: Benchmarking WebUI Code Generators with UI-Agent Execution Harness", "notes": "Adjacent evaluation infrastructure: the target is generated WebUI functionality, not a general score for computer-use agents. This annotation is based on the complete retained abstract; released artifacts and experimental results were not independently inspected or reproduced.", "published_at": "2026-10-02", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces an execution-based benchmark for generated web interfaces, with a headless-browser UI-agent harness that checks DOM interactions and functional assertions.", "topics": ["evaluation", "web-navigation", "ui-design"], "updated_at": "2026-10-02", "url": "https://arxiv.org/abs/2610.02617"}
{"aliases": [], "authors": ["Yulong Ming", "Jie Xu", "Zihan Wu", "Xiaohua Jia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "923a2e8c6f95b38f84e66e4d2cacdfd54aae8348e17b3fab858d2d7151e82d5b", "submitted_at": "2026-10-02T07:22:46Z", "updated_at": "2026-10-02T07:22:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02932v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.02932", "identifiers": {"arxiv": "2610.02932"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.02932"}], "name": "When to Compile a Computer-Use Agent? Measuring Payback and Making Compilation Decisions for Token Efficiency", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PACE, a protocol and online decision rule for deciding when repeated GUI procedures are worth compiling into programs. It accounts for failed compilation attempts, measured per-use savings, uncertain reuse, and execution costs under an explicit cumulative budget.", "topics": ["automation", "efficiency", "workflow"], "updated_at": "2026-10-02", "url": "https://arxiv.org/abs/2610.02932"}
{"aliases": [], "authors": ["Jiangang Han"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c25f023b6ff66baa38661362d5730fc338e781db5bd23dd91216d3937ae730b0", "submitted_at": "2026-10-02T09:18:28Z", "updated_at": "2026-10-07T13:46:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.03036v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.03036", "identifiers": {"arxiv": "2610.03036"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.03036"}], "name": "WebFovea: When the Model Is Right but the Click Is Wrong -- Reliable Round Trips for Vision-Based Web Agents on Live Websites", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes WebFovea, a vision-based live-web agent whose execution loop checks action parsing, browser effects, feedback accuracy, and observation delivery. The report analyzes failures such as coordinate mismatches and silent control interactions and documents corresponding harness changes.", "topics": ["web-navigation", "robustness", "observability"], "updated_at": "2026-10-07", "url": "https://arxiv.org/abs/2610.03036"}
{"aliases": [], "authors": ["Giulio Zingrillo", "Hanna Foerster", "Ilia Shumailov", "Yiren Zhao", "Robert Mullins"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7469484bb1186dd633b26e41dddc890008e893c303f93d2352f93c9d47fd78c2", "submitted_at": "2026-10-02T10:08:39Z", "updated_at": "2026-10-02T10:08:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.03089v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.03089", "identifiers": {"arxiv": "2610.03089"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.03089"}], "name": "Securing Computer-Use Agents Against Branch Steering Attacks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces STEER-Bench to study branch-steering attacks that manipulate untrusted web data toward hazardous paths already permitted by an agent’s plan. Its COBRA architecture combines trusted branching plans with advance constraints on the parameters and destinations each branch can use.", "topics": ["safety-security", "planning", "evaluation"], "updated_at": "2026-10-02", "url": "https://arxiv.org/abs/2610.03089"}
{"aliases": [], "authors": ["Alham Fikri Aji", "Faiz Rizki Ramadhan", "Zayd M. K. Zuhri", "Seung Hun Eddie Han", "Ryandito Diandaru", "Qinrong Cui", "Jan Christian Blaise Cruz", "Badrinath Chandana", "Peerawat Chomphooyod", "Ahmed Attia", "Jonibek Mansurov", "Emilio Villa-Cueva", "Canh Duong Nguyen", "Imran Turganov", "Minghao Wu", "Peerat Limkonchotiwat", "Irina Nikishina"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "31b58479f71502134f996fbdedb1dfc6fcaf4053ea8dba617a7bb538336751ec", "submitted_at": "2026-10-02T16:48:49Z", "updated_at": "2026-10-02T16:48:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.03574v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2610.03574", "identifiers": {"arxiv": "2610.03574"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.03574"}], "name": "HyperBrowseComp: A Multilingual and Multimodal Stress Test for Web-Browsing Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-02", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-02", "url": "https://arxiv.org/abs/2610.03574"}
{"aliases": [], "authors": ["Maria Sanchez del Rio", "Mikhail Terekhov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f9dd28bdfb8879cbdb68f539c286365a06f4e9d04eba2420bb46f3575fdbfa05", "submitted_at": "2026-10-02T00:31:34Z", "updated_at": "2026-10-12T23:17:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.03818v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.03818", "identifiers": {"arxiv": "2610.03818"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.03818"}], "name": "Control OSWorld: An AI Control Environment for GUI Computer Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-02", "relations": [], "review_status": "curated", "scope": "core", "summary": "Pairs OSWorld tasks with harmful side objectives to evaluate monitors that detect malicious behavior in computer-use agent trajectories.", "topics": ["safety-security", "evaluation"], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2610.03818"}
{"aliases": [], "authors": ["Yadong Lu", "Theodore Lee", "Yifei Li", "Lawrence Keunho Jang", "Tianci Xue", "Yu Su", "Huan Sun", "Ahmed Hassan Awadallah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97af38af8647b769c129d917b65f8f744035308f6ace88706a4888a96fe345b9", "submitted_at": "2026-10-02T22:43:06Z", "updated_at": "2026-10-02T22:43:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.04116v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2610.04116", "identifiers": {"arxiv": "2610.04116"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.04116"}], "name": "CUAWright: A Minimal Unified Interface for Digital Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-02", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Studies a minimal terminal-based execution framework that lets digital agents construct tools and manage context across browser, desktop, and CAD tasks.", "topics": ["tool-integration", "efficiency"], "updated_at": "2026-10-02", "url": "https://arxiv.org/abs/2610.04116"}
{"aliases": [], "authors": ["Ziyang Long", "Xinqi Li", "Lujing Xing", "Hsin-Jung Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "baacd8b4fad7b146bb9152a54d60a617fdbed5d0142c0c0768dacea2e7f18e60", "submitted_at": "2026-10-03T22:56:09Z", "updated_at": "2026-10-03T22:56:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.04800v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.04800", "identifiers": {"arxiv": "2610.04800"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.04800"}], "name": "MedImageOSWorld: Benchmarking GUI Agents for Medical Image Consoles", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-03", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates GUI agents in simulated medical-imaging consoles, separately checking workflow completion and acquisition outcomes against hidden anatomical targets.", "topics": ["evaluation", "science", "planning"], "updated_at": "2026-10-03", "url": "https://arxiv.org/abs/2610.04800"}
{"aliases": [], "authors": ["Jiazheng Zhang", "Long Ma", "Yunxian Yang", "Zhiheng Xi", "Zhikai Lei", "Yajie Yang", "Chenyang Liao", "Enyu Zhou", "Yang Nan", "Yuchen Tian", "Senjie Jin", "Yibo Wang", "Wei He", "Boyang Liu", "Jixuan Huang", "Xin Guo", "Shutong Jiang", "Zhezheng Hao", "Xinbing Liang", "Zhihao Zhang", "Changzhi Zhou", "Wiggin Zhou", "Tao Gui", "Qi Zhang", "Xuanjing Huang", "Clarenceai", "Aiden Adams"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff636b3066706043a44a69917d779771ed97e55467396679dd824836ba41387c", "submitted_at": "2026-10-04T03:36:14Z", "updated_at": "2026-10-13T09:55:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2610.04906v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2610.04906", "identifiers": {"arxiv": "2610.04906"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.04906"}], "name": "Scaling Verifiable Environments for Long-horizon Work Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.04906"}
{"aliases": [], "authors": ["Maxence Lasbordes", "Aarush Sinha", "Raphael Sourty", "Amélie Chatelain", "Djamé Seddah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5ee91448316dd9e05f89375fafe16f5998a2e3010898be4293e346958bbc0684", "submitted_at": "2026-10-04T10:13:41Z", "updated_at": "2026-10-04T10:13:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.05099v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.05099", "identifiers": {"arxiv": "2610.05099"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.05099"}], "name": "Small Agents with Semantic Search: Efficient Multilingual Code Localization", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-04", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-04", "url": "https://arxiv.org/abs/2610.05099"}
{"aliases": [], "authors": ["Kaiyu Wu", "Beichen Zheng", "Weiyao Huang", "Keze Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d20eb090847d25deb8f69181caa6f964661bbf167223b5572cf3e1cafa52bbd", "submitted_at": "2026-10-04T12:49:01Z", "updated_at": "2026-10-04T12:49:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.05185v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.05185", "identifiers": {"arxiv": "2610.05185"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.05185"}], "name": "Recurrent Latent Visual Search for GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ReLaViS, which performs recurrent visual search inside a grounding model by repeatedly querying screenshot tokens and feeding the resulting spatial evidence back into the model. Coarse-to-fine supervision is constructed from existing interface-element annotations.", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2026-10-04", "url": "https://arxiv.org/abs/2610.05185"}
{"aliases": [], "authors": ["Justin Chih-Yao Chen", "Elias Stengel-Eskin", "Yan Chen", "Pol Llado", "Scott Counts", "Mohit Bansal", "Benjamin Van Durme", "Harsh Jhamtani", "Gaurav Verma"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9e40e7a3404f7aef992270e54986e655995ae88dd0c50db55ccd7f61cc90fb3", "submitted_at": "2026-10-04T18:19:33Z", "updated_at": "2026-10-06T03:17:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.05437v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2610.05437", "identifiers": {"arxiv": "2610.05437"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.05437"}], "name": "TeleTune: Evolving Agent Skills From Offline Telemetry", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-04", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces TeleTune, which edits a textual skill library using held-out action prediction on offline software-use logs with missing goals or interleaved tasks. The learned workflows also guide retrieval of demonstrations for subsequent web-agent execution.", "topics": ["memory", "workflow", "efficiency"], "updated_at": "2026-10-06", "url": "https://arxiv.org/abs/2610.05437"}
{"aliases": [], "authors": ["Yongxin Ning", "Runliang Niu", "Qianli Xing", "Zhiyi Duan", "Qingzu He", "Pan Wang", "Qi Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43845ad64843a0b019c1988d4ba861eccee3c48a7b2c09d0bae7794e5e86347a", "submitted_at": "2026-10-05T06:28:58Z", "updated_at": "2026-10-05T06:28:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.05861v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.05861", "identifiers": {"arxiv": "2610.05861"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.05861"}], "name": "Imagine to Act: High-Fidelity Data Synthesis via Image Editing World Model for Scalable GUI Agent Training", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Infinite-Dreamer, which learns action-conditioned screenshot editing from structured descriptions of GUI changes and uses it to synthesize robustness examples and multi-step trajectories. The synthetic data trains a mobile GUI agent evaluated on AndroidWorld, MobileWorld, and AndroidControl-Curated.", "topics": ["world-models", "data-generation", "mobile-navigation"], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2610.05861"}
{"aliases": [], "authors": ["Yifan Zhang", "Yutong Dai", "Viraj Prabhu", "Zhiyuan Hu", "Ran Xu", "Zeyuan Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f3e9860574951cd6b749d47da748e4aec36d4f5afe226f4a3d0d5459fd07935", "submitted_at": "2026-10-05T17:57:51Z", "updated_at": "2026-10-05T17:57:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.06829v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.06829", "identifiers": {"arxiv": "2610.06829"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.06829"}], "name": "CLIFT: Conformal Self-Verification for Web Agent Training and Test-Time Scaling", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CLIFT, which certifies agent-generated verification signals against training-time judges and reuses them for per-step rewards and trajectory selection. At inference, a frozen question bank helps compare candidate web rollouts without querying an external judge.", "topics": ["reinforcement-learning", "evaluation", "web-navigation"], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2610.06829"}
{"aliases": [], "authors": ["Li Hu", "Kanghua Mo", "Yingbin Jin", "Qingqing Ye", "Haibo Hu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5afea9414462386ee921b9ead9381c0204963c710c9f616a96f4bc47079f059e", "submitted_at": "2026-09-29T03:03:00Z", "updated_at": "2026-09-29T03:03:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.06898v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:31.520084+00:00", "role": "primary", "sha256": "f053c4c65f6443acbda26e25cf36f7b6e8ea45f2916535604229f7c4e62ed679", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:31.520084+00:00", "role": "primary", "sha256": "f053c4c65f6443acbda26e25cf36f7b6e8ea45f2916535604229f7c4e62ed679", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:10:35.098754+00:00", "role": "primary", "sha256": "73f83c27e58e46086dbd5f32f108df2c5faa9ab2bd968d48062f39d94b2cc14b", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2610.06898v1"}], "id": "arxiv:2610.06898", "identifiers": {"arxiv": "2610.06898"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:31.520084+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.06898"}], "name": "DIBench: Benchmarking Decision Integrity of GUI-based Mobile Agents Under Deceptive Injections", "notes": "The commercial-app protocol uses fixed screenshots and context for selection decisions, rather than live end-to-end transactions. Android is supported by the paper's emulator setup. Completion and integrity are distinct measures; no reported scores were reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3.1 Benchmark components; 3.2 Instance generation; 3.3 Paired protocol/metrics; 5 Limitations/Ethics; Appendix B (Android emulator setup)); not an entire-paper reading or reproduction claim.", "published_at": "2026-09-29", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates whether deceptive UI content steers mobile agents away from instruction-consistent choices, using paired benign and injected decision states and decision-integrity metrics.", "topics": ["evaluation", "safety-security", "mobile-navigation"], "updated_at": "2026-09-29", "url": "https://arxiv.org/abs/2610.06898"}
{"aliases": [], "authors": ["Chinmay Savadikar", "Zhaoyu Zhang", "Mingyu Zhao", "Shuang Xie", "Han Li", "Tianfu Wu", "Lingyun Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4890856d5961395a9ac8c83c60b988da4adeaab1d99edd511e4046e4f99fd25d", "submitted_at": "2026-10-05T17:11:07Z", "updated_at": "2026-10-05T17:11:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07118v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.07118", "identifiers": {"arxiv": "2610.07118"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.07118"}], "name": "AMBER: Training Long-Horizon Web Agents through Append-Only Memory", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AMBER, an append-only memory framework in which a web agent learns to reason, act, and retain evidence using outcome-reward reinforcement learning. It studies whether preserving entries avoids the loss of facts and corrective feedback seen with rewritten memory.", "topics": ["memory", "long-horizon", "reinforcement-learning"], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2610.07118"}
{"aliases": [], "authors": ["Dheeraj Varghese", "Anna Vettoruzzo", "Walter Simoncini", "Michelle Lorena Acevedo Callejas", "Mohammad Mahdi Derakhshani", "Kristof Meding", "Joaquin Vanschoren", "Cees G. M. Snoek"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "21a807604326825ca2159d3528f0c12c8127043703f7fd56f0dd0213c7697d50", "submitted_at": "2026-10-05T17:52:40Z", "updated_at": "2026-10-05T17:52:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07127v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2610.07127", "identifiers": {"arxiv": "2610.07127"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.07127"}], "name": "PlaySuite: A Large-Scale Benchmark for Interactive Visual Intelligence", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-05", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2610.07127"}
{"aliases": [], "authors": ["Aadi Chauhan", "Arthur Ilyasov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81342daf25b0fb26f6481fcf836c3d958e26c2358be96eadb51e6914ea87e43c", "submitted_at": "2026-10-05T21:50:50Z", "updated_at": "2026-10-05T21:50:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07444v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.07444", "identifiers": {"arxiv": "2610.07444"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.07444"}], "name": "Decoupling What from Where: How Should a Small GUI Grounding Model Receive the Action Type?", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-05", "relations": [], "review_status": "curated", "scope": "core", "summary": "Compares ways to condition a small GUI grounding model on action type under matched training conditions on Android in the Wild. The study traces apparent gains to a coordinate-serialization choice and reports limited evidence that the gains persist with predicted action types.", "topics": ["grounding", "evaluation", "robustness"], "updated_at": "2026-10-05", "url": "https://arxiv.org/abs/2610.07444"}
{"aliases": [], "authors": ["Yeji Park", "Jaeyun Shim", "Taesik Gong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b792cc0355a0947c61014c471c482aef173f761f7396a939c14816ac644abe66", "submitted_at": "2026-10-06T08:38:30Z", "updated_at": "2026-10-06T08:38:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07972v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.07972", "identifiers": {"arxiv": "2610.07972"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.07972"}], "name": "Can Agents Work for Everyone? Cross-User Reliability for Mobile GUI Agents in Personalized User Interfaces", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PAIR for controlled evaluation of mobile GUI agents across user-conditioned application states and RePAIR for training from cross-user subgoal outcomes. The study examines failures when personalized interfaces change which items an agent must identify.", "topics": ["evaluation", "robustness", "mobile-navigation", "reinforcement-learning"], "updated_at": "2026-10-06", "url": "https://arxiv.org/abs/2610.07972"}
{"aliases": [], "authors": ["Guangyi Liu", "Yong Liu", "Jiangning Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "356c7eb45841a33a2420fbe88fd4487c72461c64da7fc308ed97c480e849617c", "submitted_at": "2026-10-06T17:13:53Z", "updated_at": "2026-10-06T17:13:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.08699v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "id": "arxiv:2610.08699", "identifiers": {"arxiv": "2610.08699"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.08699"}], "name": "nanoMuse: An Open-Source Personal Agent for Every Device You Own", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-06", "url": "https://arxiv.org/abs/2610.08699"}
{"aliases": [], "authors": ["Sarim Hashmi", "Mukul Ranjan", "Kshitij Mishra", "Mikhail Kuznetsov", "Praneeth Vepakomma", "Nils Lukas"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7113025e0dfd784d6994108c24681000625b13598e33ad0e4324e90fd2832c19", "submitted_at": "2026-10-06T17:56:43Z", "updated_at": "2026-10-06T17:56:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.08773v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.08773", "identifiers": {"arxiv": "2610.08773"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.08773"}], "name": "AdvSim2Real : Training Web Agents Against Adaptive Prompt Injection in a Web World Model", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-06", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents AdvSim2Real, which jointly adapts web-task curricula, prompt-injection adversaries, and agent policies inside a frozen web world model. It evaluates robustness against an unseen adversary and transfer from simulated interaction to a real browser.", "topics": ["safety-security", "world-models", "reinforcement-learning"], "updated_at": "2026-10-06", "url": "https://arxiv.org/abs/2610.08773"}
{"aliases": [], "authors": ["Gennadiy Savrasov", "Maksim Elistratov", "Nikita Gavrilov", "Albert Garifullin", "Oleg Pavlov", "Soslan Kabisov", "Vladimir Frolov", "Anton Konushin", "Andrey Kuznetsov", "Dmitrii Zhemchuzhnikov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d2f9c678478fe7054983e92de4dd3206f806a1dfe45e8095b3fac45173a5ed0", "submitted_at": "2026-10-06T21:20:59Z", "updated_at": "2026-10-06T21:20:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.09127v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.09127", "identifiers": {"arxiv": "2610.09127"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.09127"}], "name": "CADFather: Autonomous CAD Reconstruction through Coordinated Tool Use", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-06", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-06", "url": "https://arxiv.org/abs/2610.09127"}
{"aliases": [], "authors": ["Wanjing Han", "Levi Taiji Li", "Mu Zhang", "Yue Jiang", "Guanhong Tao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3b6ef7e785ad509ebd13fdb3da97fc583ee4c28dd42b4eab842c6cd88d7ef51f", "submitted_at": "2026-10-07T00:05:53Z", "updated_at": "2026-10-07T00:05:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.09240v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.09240", "identifiers": {"arxiv": "2610.09240"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.09240"}], "name": "Adversarial Images Hijack Web Agents from Visual Grounding to Browser Execution", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebMirage, a visual red-teaming framework that models webpage element competition and action post-processing to test whether manipulated grounding decisions lead to executed browser actions.", "topics": ["safety-security", "grounding", "evaluation"], "updated_at": "2026-10-07", "url": "https://arxiv.org/abs/2610.09240"}
{"aliases": [], "authors": ["Sarthak Choudhary", "Mihai Christodorescu", "Ashish Hooda", "Somesh Jha", "Tongxin Li", "Damien Octeau"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f38c9e37ab7fc8356a0c8c72370de673d1c2b6294e15dfdc72c280269f43439c", "submitted_at": "2026-10-07T05:24:40Z", "updated_at": "2026-10-07T05:24:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.09469v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.09469", "identifiers": {"arxiv": "2610.09469"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.09469"}], "name": "Secure-CUA: Controlling Untrusted Influence in Computer-Use Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Proposes Secure-CUA, which commits to per-action transactions before reading untrusted page content and separates content queries from GUI target selection. The paper formalizes its protection assumptions and evaluates benign task utility on WebArena.", "topics": ["safety-security", "web-navigation"], "updated_at": "2026-10-07", "url": "https://arxiv.org/abs/2610.09469"}
{"aliases": [], "authors": ["Hoang Phan", "Dat Huynh", "Andrey Zhmoginov", "Qi Zeng", "Wancen Mu", "Yue Cao", "Shengjie Bi", "Yun He", "Changdae Oh", "Deren Lei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b33a0d9dbe8705a43b4b3792fa8b2e46ac3db98e86877bab2ae58c7d47e17e3", "submitted_at": "2026-10-07T05:37:09Z", "updated_at": "2026-10-08T17:15:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.09484v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.09484", "identifiers": {"arxiv": "2610.09484"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.09484"}], "name": "MIMESIS: Learning User Simulators as Training Environments for Interactive Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2610.09484"}
{"aliases": [], "authors": ["Igor Slinko", "Yaroslav Golubev", "Sergey Titov"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15c234299d50dee8b36f8092776d0fbaeff17b037f1b57759dd1f4c883dc2d00", "submitted_at": "2026-10-07T08:08:39Z", "updated_at": "2026-10-11T11:45:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.09633v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.09633", "identifiers": {"arxiv": "2610.09633"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.09633"}], "name": "Coding-Agent Benchmarks Should Match Their Users' Task Flows", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-07", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-11", "url": "https://arxiv.org/abs/2610.09633"}
{"aliases": [], "authors": ["Michael Ofengenden", "Maksym Andriushchenko"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03aea09f3ed1b75a8a6584efdfa51d45d1cfa9dc53fb3a25efd820c5b72e094a", "submitted_at": "2026-10-07T12:26:24Z", "updated_at": "2026-10-08T09:06:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.09944v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T13:48:10Z", "role": "primary", "sha256": "6d8c0463b6d50320246021f0394d0e39c9fd839f1ef1f742c8cdfa231872b059", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:08:56.488063+00:00", "role": "primary", "sha256": "5424a5f9278ef276c84799a08d4e2e171ee93947a38c1d6670c36fa76dc84fe3", "supports": ["name", "kind", "summary", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.09944v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:15:05.368243+00:00", "role": "primary", "sha256": "9bef76f95b41a7ca39c6fa3276d6122e4b27e1e0bd56a777edcfd7e11e7d7342", "supports": ["environments", "notes"], "url": "https://github.com/michaelofengenden/agenttimebench"}], "id": "arxiv:2610.09944", "identifiers": {"arxiv": "2610.09944"}, "kind": "paper", "last_verified_at": "2026-10-20T15:08:56.488063+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.09944"}], "name": "AgentTime: Can Agents Estimate and Control Their Own Runtime?", "notes": "This cross-domain suite includes 16 OSWorld 2.0 tasks among 222 tasks from 18 sources. A single run per task and requested duration, provider effects and incomplete native grades limit comparisons; elapsed time is not evidence of productive work. The OSWorld task 092 video judge was skipped without an API key. The inspected repository README identifies desktop-VM timing and states that the current 222-task forecasting runs will be added; its existing forecasting code covers an earlier study, and transcript work/wait labels are not public. No code or benchmark was executed and no license was verified.", "published_at": "2026-10-07", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates duration following, runtime forecasts and elapsed-time estimates in native agent harnesses, including OSWorld desktop tasks, with external timers, native graders and work-versus-wait analysis.", "topics": ["evaluation", "observability"], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2610.09944"}
{"aliases": [], "authors": ["Xing Han Lù", "Dheeraj Vattikonda", "Sina Hajimiri", "Fatemeh Pesaran Zadeh", "Parishad BehnamGhader", "Ghazwa Darwiche", "Amirhossein Kazemnejad", "Christopher Pal", "Alexandre Drouin", "Siva Reddy"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37413f11bf8263f1516a8edca8c594227fcab104c0c767a7daa8eead9645ec34", "submitted_at": "2026-10-08T01:03:48Z", "updated_at": "2026-10-08T01:03:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11050v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.11050", "identifiers": {"arxiv": "2610.11050"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.11050"}], "name": "AgentHorizon: Evaluating Agentic Judges for Long-Horizon Computer-Use Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces AgentHorizon, a benchmark for judging long computer-use trajectories using closely related instructions paired with recorded actions and screenshots. It tests whether judges can detect constraint violations and distinguish successful tasks from similar but incompatible requests.", "topics": ["evaluation", "long-horizon", "cross-app"], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2610.11050"}
{"aliases": [], "authors": ["Hao Jiang", "Yiru Mao", "Tianpeng Bu", "Hao Zhou", "Hongtao Duan", "Wang Jing", "Bowen Xu", "Xin Chen", "Lulu Hu", "Bin Yang", "Yongliang Tao", "Minying Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff40a4f64eb62cf50a1c96201e89314a3355a258c5ecf6f29541155f9b4b7470", "submitted_at": "2026-10-08T04:55:50Z", "updated_at": "2026-10-08T04:55:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11251v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.11251", "identifiers": {"arxiv": "2610.11251"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.11251"}], "name": "V-CoLA: Vision Token Compression with Linear Attention", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2610.11251"}
{"aliases": [], "authors": ["Yuan Tian", "Bing Hu", "Hao Wang", "Binghang Lu", "Fang Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "506af2da01ecd6c5ee98613c634a0a5f22b50498ac32422663546a8c181f0f76", "submitted_at": "2026-10-08T10:33:29Z", "updated_at": "2026-10-08T10:33:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11655v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.11655", "identifiers": {"arxiv": "2610.11655"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.11655"}], "name": "Harness Evolution Hits a Ceiling: When Weight Training Should Begin", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2610.11655"}
{"aliases": [], "authors": ["Yudong Bai", "Yihong Chen", "Quanming Yao", "Yaqing Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2627960a5698d4c2f245e91ca5033be8c09e8a1ddea02490fcc1ed9f3d42217d", "submitted_at": "2026-10-08T11:12:57Z", "updated_at": "2026-10-08T11:12:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11707v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.11707", "identifiers": {"arxiv": "2610.11707"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.11707"}], "name": "DeltaReplay: Task-Relative Memory Reuse for Mobile GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces DeltaReplay, which stores mobile GUI trajectories as transition graphs and decides at each step whether to reuse an action, substitute task-specific parameters, or defer to the base agent. This supports reuse of partially matching experiences without rewriting stored memory.", "topics": ["memory", "mobile-navigation", "planning"], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2610.11707"}
{"aliases": [], "authors": ["Jimmy Lin", "Sahel Sharifymoghaddam", "Lingwei Gu", "Nour Jedidi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf15c34d038a85cd218337b2d86d3dc9e1b2049b97c16674b6559173f4b56245", "submitted_at": "2026-10-08T13:19:46Z", "updated_at": "2026-10-08T13:19:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11922v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.11922", "identifiers": {"arxiv": "2610.11922"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.11922"}], "name": "Project Greenhouse: Progress Toward Fully Open and Sovereign Agentic Search", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-08", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2610.11922"}
{"aliases": [], "authors": ["Jiaming Zhang", "Xuan Wang", "Fuyao Zhang", "Yang Cao", "Lingjuan Lyu", "Wei Yang Bryan Lim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53120f960c7b98e87210c93ac661a2e397b858cc9f211786b0f8233fa61c6dd3", "submitted_at": "2026-10-08T13:33:23Z", "updated_at": "2026-10-08T13:33:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11942v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.11942", "identifiers": {"arxiv": "2610.11942"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.11942"}], "name": "Right Screen, Wrong Transition: World Models as Verifiers for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents LGWM, a decoder-free GUI world model that predicts the next screen representation from an action and compares it with the observed state to verify transitions. Its paired diagnostic separates legitimate and hijacked transitions that lead to the same credential screen.", "topics": ["world-models", "safety-security", "evaluation"], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2610.11942"}
{"aliases": [], "authors": ["Utku Boran Torun", "Veli Karakaya", "Eray Tüzün"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e77dcf101ba95684aac51084e43b914c0aca86fc25d061612224118004bdb21", "submitted_at": "2026-10-08T13:44:53Z", "updated_at": "2026-10-08T13:44:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11963v1"}}, "environments": ["web", "desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T12:58:20.217517Z", "role": "primary", "sha256": "f275c47a8b6d349a70e39a9e55eea5a09d99d5f35cc896100f3bcfd5b0923e0b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202609200000+TO+202610201255%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:25.581656+00:00", "role": "primary", "sha256": "9fad44164a6f00446d2c91a31e3e77a4aab8145ac3747e9d1ce0165c1aebda95", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.11963v1"}], "id": "arxiv:2610.11963", "identifiers": {"arxiv": "2610.11963"}, "kind": "paper", "last_verified_at": "2026-10-20T13:12:25.581656+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.11963"}], "name": "Can LLMs Fix It Without Code? Toward Automated Verification of No-Code Bug Fixes", "notes": "The evaluation reuses 322 fixes from 12 generation configurations and is restricted to nine of 51 candidate issues after environment setup and sanity checks. Resolution counts any success in three runs with a 20-step episode limit; executor harnesses and decoding settings differ. Opening the update page triggers a harness-controlled switch to the fixed browser build. Checker behavior, reconstructed starting states and the two-author human reference limit interpretation. The replication package was linked by the paper but was not inspected; no software, agent, model or result was run or reproduced.", "published_at": "2026-10-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Tests natural-language no-code fixes through GUI execution in Linux Brave. Scripted fixed/buggy-state gates and issue-specific checkers distinguish fix outcomes from the executor's completion claims.", "topics": ["evaluation", "web-navigation"], "updated_at": "2026-10-08", "url": "https://arxiv.org/abs/2610.11963"}
{"aliases": [], "authors": ["Tian Jin", "Wenhao Wang", "Jing Kang", "Haoting Shi", "Rui Ye", "Yijun Lu", "Siheng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "06b25b25e804683e11305e21dd04feb9131aad27f901998eda70ac91e527c6b9", "submitted_at": "2026-10-08T18:31:16Z", "updated_at": "2026-10-14T08:40:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.12611v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.12611", "identifiers": {"arxiv": "2610.12611"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.12611"}], "name": "SoftwareGym: Scaling Computer-Use Agent Training with Diverse Synthesized Software Environments", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-08", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SoftwareGym, a pipeline that synthesizes runnable GUI applications and iteratively repairs them using execution checks. It also provides a task benchmark and studies whether training on synthesized application trajectories transfers to other computer-use environments.", "topics": ["data-generation", "evaluation", "sandbox"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.12611"}
{"aliases": [], "authors": ["Yisheng Li", "Dongming Han", "Hongkun Pan", "Bingcheng Mao", "Zijian Wang", "Yunlong Mao", "Wei Tong", "Sheng Zhong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6f1521fdfdf6a11a2a195feb05fd915afd6e9bb2332486b870da6375214ac1a", "submitted_at": "2026-10-09T00:03:13Z", "updated_at": "2026-10-09T00:03:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.12823v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.12823", "identifiers": {"arxiv": "2610.12823"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.12823"}], "name": "When Charts Deceive Web Agents: A Paired Benchmark for Misleading Visualization-Grounded Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MisVisAgentBench, a paired executable benchmark that changes chart evidence while holding web tasks fixed, to trace how misleading visualizations affect agent decisions and downstream actions.", "topics": ["evaluation", "robustness", "screen-understanding"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.12823"}
{"aliases": [], "authors": ["Shangquan Sun", "Ruoyu Chen", "Zhirui Liu", "Xiang Liu", "Chenghao Qian", "Wenqi Ren", "Xiaochun Cao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f7effcf5e101cfcc7e229fe4aab40d07401b1dd98cf5ce3b566fef863a42190", "submitted_at": "2026-10-09T01:50:26Z", "updated_at": "2026-10-09T01:50:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.12866v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.12866", "identifiers": {"arxiv": "2610.12866"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.12866"}], "name": "CaptchaAgent: Verifiable Visual Reasoning for CAPTCHA Solving", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents CaptchaAgent, which builds an explicit visual representation of objects and relations, refines ambiguous regions, and checks counterfactual consistency before acting on interactive CAPTCHAs.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.12866"}
{"aliases": [], "authors": ["Yuqi Huang", "Hui Li", "Yifei Cao", "Zikun Cai", "Tianci Xie", "Yangfan Gao", "Pinxue Guo", "Chaoyun Zhang", "Bingxuan Li", "Rui Zheng", "Hang Yan", "Ning Liao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e90bb713e7034a64821502eb6fc792fa91a5d1e420aff9ff658506d39269f03", "submitted_at": "2026-10-09T04:32:02Z", "updated_at": "2026-10-09T04:32:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.12984v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.12984", "identifiers": {"arxiv": "2610.12984"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.12984"}], "name": "Work2Skill: Capturing Human GUI Experience for Reusable Agent Skill Induction", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Aligns human screen recordings, input events, and accessibility information to derive reusable workflow skills from desktop demonstrations.", "topics": ["recording", "workflow", "data-generation"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.12984"}
{"aliases": [], "authors": ["Yuchen Ling", "Shengcheng Yu", "Zhenyu Chen", "Chunrong Fang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb5f0220266a43b74e7a5f547abbe78a54a2ef993fee5d27f67827cb9715f5dc", "submitted_at": "2026-10-09T06:45:23Z", "updated_at": "2026-10-09T06:45:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13092v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:08:54Z", "role": "primary", "sha256": "3db7474953cbe5a3052fb22793038f2383bdd0e7c775af06494a5e8501646d18", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.13092v1"}], "id": "arxiv:2610.13092", "identifiers": {"arxiv": "2610.13092"}, "kind": "paper", "last_verified_at": "2026-10-20T15:08:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.13092"}], "name": "When Faster Decisions Do Not Make Better Agents: An Empirical Study of Jev in Agent Workflows", "notes": "The mobile evaluation uses text UI observations and one fixed instance for each of 116 AndroidWorld task types; three repeats share seed 0. Five agent families use different budgets, with candidate-interface and tool-call granularity changes in some variants. Reported costs use fixed experimental prices, and outcome-selected trace labels are not population failure rates. Faster selection does not imply faster or more reliable task completion. Selected sections were source-read; no independent reproduction.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Evaluates Jev as a bounded action selector inside three AndroidWorld agents and two text-based ALFWorld agents, separating decision savings from task success, recovery and resources per delivered outcome.", "topics": ["efficiency", "action-prediction", "evaluation", "mobile-navigation", "self-reflection"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.13092"}
{"aliases": [], "authors": ["Yingbo Ma", "Kuangshi Ai", "Gordon Dai", "Chenglong Yu", "Chaoli Wang", "Hua Shen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "35908311612dc14e8ff82888255e962b6e6a86349117d67e7d6a77758a6d8e16", "submitted_at": "2026-10-09T06:51:59Z", "updated_at": "2026-10-12T20:16:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.13098v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.13098", "identifiers": {"arxiv": "2610.13098"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.13098"}], "name": "VETR: Execution-Grounded Self-Improving Agent for Symbolic Graphics Reasoning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2610.13098"}
{"aliases": [], "authors": ["Shun Zou", "Ziyu Ma", "Yi Zou", "Yong Wang", "Lin Chen", "Zehui Chen", "Guanhua Chen", "Xiangxiang Chu", "Feng Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f6fbb06b6a6f39f8c823e53974d2be3d37d11e0b121adb6bafde8a25de9c717e", "submitted_at": "2026-10-09T07:55:55Z", "updated_at": "2026-10-09T07:55:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13166v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.13166", "identifiers": {"arxiv": "2610.13166"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.13166"}], "name": "Scaling Agentic Data for Long-Horizon Terminal Intelligence", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.13166"}
{"aliases": [], "authors": ["Ruotian Zhang", "Yuhui Zhang", "Junzhe Huang", "Yu Xin", "Hehe Fan", "Xiaoxiao Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "010e8ebc657dccad670ea6d17985e9c3c1680017f2516eaa86b7ea9db5660990", "submitted_at": "2026-10-09T09:20:02Z", "updated_at": "2026-10-09T09:20:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13268v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.13268", "identifiers": {"arxiv": "2610.13268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.13268"}], "name": "ConClick: Element-Wise Contrastive Learning Improves GUI Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Analyzes recurring distractor clicks in GUI grounding and introduces ConClick, a contrastive training method built from target–distractor pairs. Its ReClick inference procedure revisits candidate regions and compares competing predictions.", "topics": ["grounding", "data-generation", "robustness"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.13268"}
{"aliases": [], "authors": ["Jiahao Wang", "Hongjin Qian", "Yuyang Hu", "Jiajun Zhang", "Zheng Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea3ef3498414560547083b4ccdfcbd2d390b2b2fe59d643def9ca2e1d032575d", "submitted_at": "2026-10-09T10:35:27Z", "updated_at": "2026-10-09T10:35:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13361v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.13361", "identifiers": {"arxiv": "2610.13361"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.13361"}], "name": "AgentSWE: Can Coding Agents Build the Agent You Actually Want?", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.13361"}
{"aliases": [], "authors": ["Sunghwan Kim", "Deogyong Kim", "Sangam Lee", "Donguk Kwon", "Dongha Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cadcf2ab7be51b17d06975eba0e56573c333dca9ec034ade67b84e62d0096903", "submitted_at": "2026-10-09T11:23:40Z", "updated_at": "2026-10-09T11:23:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13416v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.13416", "identifiers": {"arxiv": "2610.13416"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.13416"}], "name": "Web$hell: A Unified Interface for Agentic Web Access", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Web$hell, an interface that compiles live pages into persistent information and action objects so agents can gather evidence, navigate websites, and reuse acquired content through shared operations.", "topics": ["web-navigation", "tool-integration", "memory", "efficiency"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.13416"}
{"aliases": [], "authors": ["Ziyi Yang", "Adel Bibi", "Guohao Li", "Bernard Ghanem", "Carlos Hinojosa"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "58dd9c59962bd6feb983bb89a66a83e5d4243292e02dcb16e436a6949e858209", "submitted_at": "2026-10-09T14:05:47Z", "updated_at": "2026-10-09T14:05:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13623v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.13623", "identifiers": {"arxiv": "2610.13623"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.13623"}], "name": "APEX-Harm: Evaluating Agent Safety in Shared Professional Workspaces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.13623"}
{"aliases": [], "authors": ["Vagner Figueredo de Santana", "Diogo Cortiz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "447a4f283746069b2ea51d49c149379ad48a6e0b71f35dd15dedcb19417ab0cd", "submitted_at": "2026-10-09T14:59:24Z", "updated_at": "2026-10-09T14:59:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13703v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.13703", "identifiers": {"arxiv": "2610.13703"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.13703"}], "name": "A Baseline Assessment for the Agentic Web Technology Stack", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents MASTIF, a framework for comparing combinations of agent models, protocols, and implementation frameworks on web tasks, with an initial assessment using 300 Mind2Web tasks.", "topics": ["evaluation", "web-navigation"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.13703"}
{"aliases": [], "authors": ["Chenyang Yan", "Yingying Zhang", "Kanzhi Cheng", "Qiushi Sun", "Zhengyuan Pan", "Zheng Ma", "Hang Yan", "Yian Wang", "Nuo Chen", "Jialin Cao", "Xingdong Gong", "Wenpo Song", "Han Chen", "Zichen Ding", "Fangzhi Xu", "Shujian Huang", "Xinyu Dai", "Yichen Liu", "Tiankuo Yao", "Bo Wang", "Ben Kao", "Jianbing Zhang", "Lewei Lu", "Dahua Lin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f7d940d3e341fd8c9f23a4829a5ab1150fc07ee231ccda9b108a3f0d785efb9", "submitted_at": "2026-10-09T17:26:38Z", "updated_at": "2026-10-09T17:26:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.14157v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:06Z", "role": "primary", "sha256": "9b385e9f6aa2f7169f616a76887df3b76df605cb2bea071d99ee7ca3f9b659cc", "supports": ["environments"], "url": "https://os-copilot.github.io/OpenMobile2-Home/"}], "id": "arxiv:2610.14157", "identifiers": {"arxiv": "2610.14157"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.14157"}], "name": "OpenMobile-2: Building Versatile Mobile Agents with Scalable Environments and App-Native Tools", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces OpenMobile-2, combining simulated mobile apps, emulator tasks, released training data, and policies that can alternate GUI interaction with app-native tools.", "topics": ["mobile-navigation", "tool-integration", "data-generation", "reinforcement-learning"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.14157"}
{"aliases": [], "authors": ["Seungjun Yu", "Hojun Choi", "Seojeong Park", "Jaeyo Shin", "Hyunjung Shim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "191e88c19124ddaddc2fed2f0e3f8e197a4c0c7f8adbc3997c056d3333754042", "submitted_at": "2026-10-09T18:45:11Z", "updated_at": "2026-10-09T18:45:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.14228v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.14228", "identifiers": {"arxiv": "2610.14228"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.14228"}], "name": "Latent World Action Model for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-09", "relations": [], "review_status": "curated", "scope": "core", "summary": "Replaces future-screenshot generation with a shared latent representation for GUI action selection and action-conditioned transition prediction. An action-aware contrastive objective encourages the transition representation to retain discriminative action-induced changes.", "topics": ["world-models", "planning", "efficiency"], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.14228"}
{"aliases": [], "authors": ["Brian Hempel", "Ruanqianqian Huang", "Devamardeep Hayatpur", "Sorin Lerner", "Haijun Xia"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b7c0f27063df5eaa6ed57cc160a0ddefff1bb5d010de9d9efa057fe925fbb6a6", "submitted_at": "2026-10-09T20:32:04Z", "updated_at": "2026-10-09T20:32:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.14306v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T13:48:10Z", "role": "primary", "sha256": "6d8c0463b6d50320246021f0394d0e39c9fd839f1ef1f742c8cdfa231872b059", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.14306", "identifiers": {"arxiv": "2610.14306"}, "kind": "paper", "last_verified_at": "2026-10-20T13:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.14306"}], "name": "Plottery: A Code-Integrated GUI for Matplotlib", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-09", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-09", "url": "https://arxiv.org/abs/2610.14306"}
{"aliases": [], "authors": ["Haowen Wan", "Shangzhuo Xie", "Syue-Ju Syue", "Xianyue Wu", "Runhua Li", "Qianqian Yang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49c5ae7f6cb9e89a0e602f5c51bc754e39434d8c8588d46449626337a5e2da24", "submitted_at": "2026-10-10T12:21:05Z", "updated_at": "2026-10-10T12:21:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.14855v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T13:48:10Z", "role": "primary", "sha256": "6d8c0463b6d50320246021f0394d0e39c9fd839f1ef1f742c8cdfa231872b059", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.14855", "identifiers": {"arxiv": "2610.14855"}, "kind": "paper", "last_verified_at": "2026-10-20T13:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.14855"}], "name": "PEAM: On-device Communication Agent for Personalized, Efficient and Autonomous Mobile Network Configuration", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-10", "url": "https://arxiv.org/abs/2610.14855"}
{"aliases": [], "authors": ["Yuxuan Liu", "Zhaochen Su", "Hangyu Guo", "Shijue Huang", "Zhixiong Zhang", "Kawai Chung", "Zhongwei Xie", "Tianshi Zheng", "Wei Fan", "Qing Zong", "Haochen Shi", "Baixuan Xu", "Haoran Li", "Yangqiu Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5891d4fcb3ec4ecc58090ccd56bd03b9863ba67874de6a460940bc864966a5fd", "submitted_at": "2026-10-10T14:36:39Z", "updated_at": "2026-10-10T14:36:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.14966v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T13:48:11Z", "role": "primary", "sha256": "89e110b6fa882639fb73936cb27e5373d11b632938c31968f563152534bf5670", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.14966%2C2610.23986&max_results=2"}], "id": "arxiv:2610.14966", "identifiers": {"arxiv": "2610.14966"}, "kind": "paper", "last_verified_at": "2026-10-20T13:48:11Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.14966"}], "name": "SkillHorizon: Constructing Long-Horizon Agent Tasks through Skill Composition", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-10", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-10", "url": "https://arxiv.org/abs/2610.14966"}
{"aliases": [], "authors": ["Md Imran Hossen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63e19265f1df2c336e28c4a13e11331021bd60866269c0b921d92a3c0fec080d", "submitted_at": "2026-10-10T19:30:31Z", "updated_at": "2026-10-10T19:30:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15138v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.15138", "identifiers": {"arxiv": "2610.15138"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.15138"}], "name": "The Agent Clicks Approve: Injected Page Text Makes Browser Agents Authorize Access the Task Never Named", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-10", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies prompt-injected account-access approvals in browser agents and proposes a service-side authority envelope that checks authorization requests against the user task before allowing them to complete.", "topics": ["safety-security", "evaluation"], "updated_at": "2026-10-10", "url": "https://arxiv.org/abs/2610.15138"}
{"aliases": [], "authors": ["Hyobin Ahn", "Dongjae Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff5470e792d14ced2cdff648d97e2606eea4b3c822046b7c10272fd2a5f1695f", "submitted_at": "2026-10-11T10:09:44Z", "updated_at": "2026-10-11T10:09:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15502v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T12:56:03.829575Z", "role": "primary", "sha256": "a8636ac679d2110181d4417d4aef040ddb24464597a7fa6fe6d2c35a2401a3b1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22graphical+user+interface%22+OR+ti%3A%22screen+grounding%22+OR+ti%3A%22UI+grounding%22+OR+ti%3A%22desktop+agents%22+OR+ti%3A%22Android+agent%22+OR+ti%3A%22phone+agent%22+OR+ti%3A%22mobile+GUI%22+OR+ti%3A%22web+navigation%22+OR+ti%3A%22GUI+automation%22+OR+ti%3A%22browser+automation%22%29+AND+submittedDate%3A%5B202608210000+TO+202610201255%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:58.069698Z", "role": "primary", "sha256": "a38b3ddc9d5d6744eefc937f86e174917f9b5de149323632d5d1b20283d79c8e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.15502v1"}], "id": "arxiv:2610.15502", "identifiers": {"arxiv": "2610.15502"}, "kind": "paper", "last_verified_at": "2026-10-20T13:12:58.069698Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.15502"}], "name": "From Observations to Authorized Actions: Measuring Source Information in Android Agents", "notes": "The two prototypes use different workloads and trust assumptions on patched Android/WebView components. Approval is simulated; authenticated browser delivery and atomic capture remain unimplemented. Coarse source history can withhold benign work and persist across tasks. Direct-call checks and limited live-agent trials are separate, and writes are in-memory. No production vulnerability, general defense rate or released artifact is established.", "published_at": "2026-10-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Measures source-attribution limits in Android accessibility and separately tests application-history and selected-WebView-object checks for authorizing AppFunctions actions.", "topics": ["safety-security", "tool-integration", "observability"], "updated_at": "2026-10-11", "url": "https://arxiv.org/abs/2610.15502"}
{"aliases": [], "authors": ["Jianlin Chen", "Wenhui Chen", "Chi Man Vong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6c60bb4d3b7bb2e2cba92201924d8da4f5e18ce5553c81fc9d5c6fb62b84751", "submitted_at": "2026-10-11T16:18:11Z", "updated_at": "2026-10-11T16:18:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15738v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.15738v1"], "reason": "Publisher document declaration only; no code, data or weights license inferred."}}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T13:48:10Z", "role": "primary", "sha256": "6d8c0463b6d50320246021f0394d0e39c9fd839f1ef1f742c8cdfa231872b059", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:31:34.620763Z", "role": "primary", "sha256": "7ec947ceb462b5fd23a4a67df64e931e427224ef721b681d3fc92c1d4379449f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.15738v1"}], "id": "arxiv:2610.15738", "identifiers": {"arxiv": "2610.15738"}, "kind": "paper", "last_verified_at": "2026-10-20T15:31:34.620763Z", "licenses": {"code": null, "data": null, "document": {"basis": "The retained version-specific arXiv HTML infobox explicitly declares License: CC BY 4.0 for this document.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T15:31:34.620763Z", "url": "https://arxiv.org/html/2610.15738v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.15738"}], "name": "Visible but Not Avoided: State-Dependent Re-Proposal of Rejected Candidates in Catalog-Based LLM Optimization Loops", "notes": "The downstream test uses one screenshot-only agent, nine synthetic apps and 60 cached development items, with TAP-tag or text-answer outputs; this is not demonstrated live-app completion. Two catalogs reword the same 40 edits. Re-proposal effects do not establish internal causal mechanisms or task-success gains. Main serving providers were unpinned; rationale labels came from LLM annotators. No independent reproduction.", "published_at": "2026-10-11", "relations": [], "review_status": "curated", "scope": "core", "summary": "Studies repeated rejected edits while optimizing a screenshot-only GUI agent, contrasting static probes with reached loop states and testing how rejection wording changes subsequent edit proposals.", "topics": ["evaluation", "grounding", "training-free"], "updated_at": "2026-10-11", "url": "https://arxiv.org/abs/2610.15738"}
{"aliases": [], "authors": ["Alan Wu", "Wei Tao", "Yiwen Guo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ef6fd8890648ae958340325edd3604fb78aa82510145169eb85f76c322b76d6", "submitted_at": "2026-10-11T16:43:43Z", "updated_at": "2026-10-11T16:43:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15765v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.15765", "identifiers": {"arxiv": "2610.15765"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.15765"}], "name": "A Decade in a Day: Synthesizing Agent-Native Interfaces from Human-Oriented Software Evolution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-11", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-11", "url": "https://arxiv.org/abs/2610.15765"}
{"aliases": [], "authors": ["Chen Huang", "Yiyang Zhang", "Yuhao Wu", "Yang Deng", "Wenxuan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c47d367230b561de34cafe047b38696279ff351904983043d07972259181f798", "submitted_at": "2026-10-12T03:31:54Z", "updated_at": "2026-10-14T06:05:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.16117v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.16117", "identifiers": {"arxiv": "2610.16117"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.16117"}], "name": "Beyond Perfect Teammates: Multi-Agent Collaboration with Unreliable Colleagues", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.16117"}
{"aliases": [], "authors": ["Jingchao Wang", "Dingjiang Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa0e0ab6eb19ea55e65ccb354ae515f6552cd03dd688c6f43cdc6e2d59b11824", "submitted_at": "2026-10-12T06:09:02Z", "updated_at": "2026-10-12T06:09:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16234v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.16234", "identifiers": {"arxiv": "2610.16234"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.16234"}], "name": "Seg2Click: Learning Pixel-Level Action Masks for GUI Visual Grounding", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Recasts GUI grounding as predicting a mask of valid click locations from bounding-box supervision, then derives a click point with a deterministic selection rule.", "topics": ["grounding", "screen-understanding"], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2610.16234"}
{"aliases": [], "authors": ["Mingwei Fu", "Zheshu Song", "Zhikai Zhou", "Shen Huang", "Shidong Shang", "Mengxiao Zhu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3256b0858be826f8058f4ed714c798583421644067d388586ef89b38fd912192", "submitted_at": "2026-10-12T06:36:20Z", "updated_at": "2026-10-12T06:36:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16260v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.16260", "identifiers": {"arxiv": "2610.16260"}, "kind": "paper", "last_verified_at": "2026-10-20T15:06:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.16260"}], "name": "ScreenASR-Bench: A Tiered Benchmark for Speech Recognition with Visual Context", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2610.16260"}
{"aliases": [], "authors": ["Zijian Cai", "Haowei Liu", "Kailai Chen", "Lin Peng", "Cong Wan", "Ziqi Zhang", "Chunfeng Yuan", "Bing Li", "Weiming Hu", "Haiyang Xu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4596cc60436c4fa6c2aa0951b4699b804edb73aa1b1a46900b1d14f901d20f3b", "submitted_at": "2026-10-12T07:42:11Z", "updated_at": "2026-10-12T07:42:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16319v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.16319", "identifiers": {"arxiv": "2610.16319"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.16319"}], "name": "HyWe: Unifying Search and Operation for Real-World Hybrid Web Tasks", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces HyWe, which switches between information search and browser operation within one trajectory, alongside a demonstration pipeline and a benchmark of hybrid web tasks.", "topics": ["web-navigation", "planning", "data-generation"], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2610.16319"}
{"aliases": [], "authors": ["Jiahao Lu", "Yize Shen", "Chuanyuan Tan", "Mianqiu Huang", "Shicheng Fang", "Xinchi Chen", "Yining Zheng", "Yuxin Wang", "Xipeng Qiu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73073d2d533307cd71b05afef2441a8a7b8f198f804f11f83536d4297db2491a", "submitted_at": "2026-10-12T12:52:21Z", "updated_at": "2026-10-12T12:52:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16720v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.16720", "identifiers": {"arxiv": "2610.16720"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.16720"}], "name": "Learn from Move: the Next Step for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces LatentGUIWorld to test learning interface semantics and dynamics through interaction, and studies finer movement actions with intermediate visual feedback and exploration demonstrations.", "topics": ["evaluation", "grounding", "reinforcement-learning"], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2610.16720"}
{"aliases": [], "authors": ["Zian Yuan", "Wei Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5a2743a18bae1deeed13fbd1dd7379c3f21c67488627057ccb466ffe51d09dc", "submitted_at": "2026-10-12T13:34:52Z", "updated_at": "2026-10-12T13:34:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16778v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.16778", "identifiers": {"arxiv": "2610.16778"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.16778"}], "name": "PrivacyGuard: Source-Linked Sanitization and Locally Resolvable PII References for Cloud GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces PrivacyGuard, which sanitizes source-linked personal information and maintains locally resolvable references across mobile-agent turns to reduce exposure to remote planners.", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2610.16778"}
{"aliases": [], "authors": ["Yuetian Chen", "Kaiyuan Zhang", "Yuntao Du", "Thomas Li", "Mark Tenenholtz", "Kyle Polley", "Jerry Ma", "Denis Yarats", "Ninghui Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "247cd4c9634dea78780a2230138c728adf53871e8b65614697f76ba53dea858c", "submitted_at": "2026-10-12T22:17:51Z", "updated_at": "2026-10-12T22:17:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17457v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.17457", "identifiers": {"arxiv": "2610.17457"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17457"}], "name": "BrowseSafe++: Benchmarking Prompt Injection Detection Across Web Contexts, Languages, and Browser Views", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-12", "relations": [], "review_status": "curated", "scope": "core", "summary": "Extends BrowseSafe with paired attacked and clean webpages in several representations, and studies text and screenshot detectors plus feedback-adapted prompt injections.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2026-10-12", "url": "https://arxiv.org/abs/2610.17457"}
{"aliases": [], "authors": ["Zirui Shang", "Yang Liu", "Xin Shu", "Zhi Gao", "Xinxiao Wu", "Lifeng Fan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "46ea5a1a350ccfcc12d22ab3783e05e3c580719962831561b01aeb3fcc6f6c28", "submitted_at": "2026-10-13T04:37:18Z", "updated_at": "2026-10-13T04:37:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17710v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.17710", "identifiers": {"arxiv": "2610.17710"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17710"}], "name": "Adaptive-Group Process Rewards for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Groups GUI trajectories by subgoal and combines action-level and group-level assessments into intermediate rewards for training agents with sparse task outcomes.", "topics": ["reinforcement-learning", "long-horizon", "planning"], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.17710"}
{"aliases": [], "authors": ["Minghua Chen", "Xingming Liao", "Weiqiang Zhong", "Zhuowei Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b1786bba4f48ca54166d49e3d98d55aaef52eb6e2233f282d8d94d54dfb5ed2c", "submitted_at": "2026-10-13T06:20:38Z", "updated_at": "2026-10-13T06:20:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17792v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.17792", "identifiers": {"arxiv": "2610.17792"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17792"}], "name": "MHVF: Training-Free GUI Grounding via Verification-Guided Multi-Hypothesis Search", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces MHVF, a training-free grounding method that explores several candidate screen regions, verifies them in local views, and uses those checks to guide further search.", "topics": ["grounding", "screen-understanding", "robustness"], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.17792"}
{"aliases": [], "authors": ["Pengcheng Huang", "Zhenghao Liu", "Shixin Ren", "Shi Yu", "Sijia Yao", "Yukun Yan", "Chaojun Xiao", "Maosong Sun"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bad610c0ceb1b551c3485d3ab797967a51b800d9ff7ac79a9808def7c0e4a7af", "submitted_at": "2026-10-13T06:40:22Z", "updated_at": "2026-10-18T09:06:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.17816v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T04:48:34Z", "role": "primary", "sha256": "4b1219a8d3b34b9ecfabab989eed959ab7e2263500ab2034fc56105db1c559eb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.22673%2C2610.17816%2C2610.20597%2C2610.23857%2C2610.25601&max_results=5"}], "id": "arxiv:2610.17816", "identifiers": {"arxiv": "2610.17816"}, "kind": "paper", "last_verified_at": "2026-10-21T04:48:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.17816"}], "name": "Empowering Edge Agents: A Systematic Analysis of Edge-Cloud Collaboration in LLM Agent Execution", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-18", "url": "https://arxiv.org/abs/2610.17816"}
{"aliases": [], "authors": ["Wayne Chi", "I-Hung Hsu", "Li Chen", "Wei-Lin Chiang", "Ion Stoica", "Anastasios Nikolas Angelopoulos", "Ameet Talwalkar", "Chris Donahue", "Cho-Jui Hsieh"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14d2128155d1540955e54c3bdf9010724f27cf2e101079650315602856c98efa", "submitted_at": "2026-10-13T07:49:50Z", "updated_at": "2026-10-14T06:37:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.17895v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.17895", "identifiers": {"arxiv": "2610.17895"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17895"}], "name": "Interaction-as-a-Judge", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses computer-use agents to interact with generated applications, studying whether interaction improves preference judging, requirement checks, and feedback for application generation.", "topics": ["evaluation", "data-generation", "automation"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.17895"}
{"aliases": [], "authors": ["Taeyun Roh", "Suhyeong Park", "Gijin Kim", "Wonjune Jang", "Jaewoo Kang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "551f198c0976c75693c6298400800cc66ce1b1b5ee304965155a6bb3ca9d68fe", "submitted_at": "2026-10-13T08:15:52Z", "updated_at": "2026-10-13T08:15:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17928v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:05Z", "role": "primary", "sha256": "3fb5aabfb07df620955f0b74347f9d5926a1cd3d506b1098eccba9477627269f", "supports": ["environments", "notes"], "url": "https://raw.githubusercontent.com/dmis-lab/SwitchMem/HEAD/README.md"}], "id": "arxiv:2610.17928", "identifiers": {"arxiv": "2610.17928"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17928"}], "name": "Learning How to Remember: Memory Modality Selection for GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. Android platform tagging is additionally supported by the primary project documentation.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SwitchMem, which chooses a textual or visual representation of retrieved experience at each GUI interaction step instead of fixing one memory modality for a task.", "topics": ["memory", "mobile-navigation"], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.17928"}
{"aliases": [], "authors": ["Guanting Dong", "Xinyu Yang", "Jiajie Jin", "Yuyang Hu", "Xiaoshuai Song", "Yifei Chen", "Xiaoxi Li", "Jiejun Tan", "Yutao Zhu", "Tat-Seng Chua", "Ji-Rong Wen", "Zhicheng Dou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3227c9f8be9728ec1c6295bb5a1196d894cfa2a4b0c11bf5c2af5e36587eb410", "submitted_at": "2026-10-13T08:36:20Z", "updated_at": "2026-10-13T08:36:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17953v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.17953", "identifiers": {"arxiv": "2610.17953"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.17953"}], "name": "Agent-Reflex: Teaching Language Agents to Act Reflectively", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.17953"}
{"aliases": [], "authors": ["Hang He", "Linting Wang", "Guanming Liu", "Bo Mao", "Jinxiu Liu", "Yiran Suo", "Maosen Zhou", "Zhiheng Xi", "Guanting Dong", "Zhou Shao", "Rui Zheng", "Jie Zhou", "Chengcheng Wan", "Tao Gui", "Qi Zhang", "Xipeng Qiu", "Xuanjing Huang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6add363f25f87ea1814c7b79ebf6b51f7431b3745a7d2d910960b78e2d97d13f", "submitted_at": "2026-10-13T09:52:02Z", "updated_at": "2026-10-13T09:52:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.18067v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.18067v1"], "reason": "Document declaration only; code, data and weights permissions are not inferred."}}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T19:02:20Z", "role": "primary", "sha256": "f1fbdd2ecb2a54b7723372118e2b0ac818420f3a3dd280c2486c8e064cf486d6", "supports": ["name", "kind", "summary", "scope", "environments", "topics", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.18067v1"}], "id": "arxiv:2610.18067", "identifiers": {"arxiv": "2610.18067"}, "kind": "paper", "last_verified_at": "2026-10-20T19:02:20Z", "licenses": {"code": null, "data": null, "document": {"basis": "The exact-v1 arXiv HTML infobox explicitly declares CC BY 4.0 for this paper.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T19:02:20Z", "url": "https://arxiv.org/html/2610.18067v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.18067"}], "name": "WorkRover: Scaling Long-Horizon Cowork Agent Training Via Recursive Environment Evolution", "notes": "App WorkEnvs expose UI observations and interaction interfaces alongside file and tool workflows; no specific operating system or pixel-control success is established. Verifier shaping depends on deterministic-check and dependency coverage. External reference scores use differing harnesses or subsets; GDPval Elo is not comparable to percentage metrics. A successful-trace analysis subset is separate from online RL outcomes. No implementation release or independent reproduction was verified.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Synthesizes resettable workspaces with files, application observations and skills; repairs environment incompatibilities and trains long-task agents using verifier-linked turn credit.", "topics": ["data-generation", "long-horizon", "evaluation", "reward-models", "reinforcement-learning", "workflow"], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.18067"}
{"aliases": [], "authors": ["Dominik Seip", "Bartol Markovinović", "Matthias Hein"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5565d38dd40b02b4158d7edcc2c5c27e7e285dbedefe1c7101a31c4f8d17b659", "submitted_at": "2026-10-13T15:03:33Z", "updated_at": "2026-10-13T15:03:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.18508v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.18508", "identifiers": {"arxiv": "2610.18508"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.18508"}], "name": "Towards Safe Autonomy for Computer Use", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents SA-CUA, separating trusted and untrusted visual content using accessibility-derived masking rules and a controlled channel for verifying information before endorsement.", "topics": ["safety-security", "screen-understanding"], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.18508"}
{"aliases": [], "authors": ["Yubo Li", "Ramayya Krishnan", "Rema Padman"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "070f02f1ea798369817f471df5fe1d002b68d544135a1ac8e3a2aed2830e8c01", "submitted_at": "2026-10-13T15:43:30Z", "updated_at": "2026-10-13T15:43:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.18579v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.18579v1"], "reason": "The retained version-specific arXiv HTML infobox explicitly labels this paper License: CC BY 4.0. This document declaration does not establish code, training-data or model-weight licenses."}}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:08:55Z", "role": "primary", "sha256": "bcbc640d1e732ba0415a9755bce120a811a13a2ba7fba3e58a167b3a9111c2a6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.18579v1"}], "id": "arxiv:2610.18579", "identifiers": {"arxiv": "2610.18579"}, "kind": "paper", "last_verified_at": "2026-10-20T15:08:55Z", "licenses": {"code": null, "data": null, "document": {"basis": "The retained version-specific arXiv HTML infobox explicitly labels this paper License: CC BY 4.0. This document declaration does not establish code, training-data or model-weight licenses.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T15:08:55Z", "url": "https://arxiv.org/html/2610.18579v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.18579"}], "name": "Reading Right, Acting Wrong: Diagnosing Decision Failures of System One Models in Agent Loops", "notes": "Adjacent diagnostic research across simulators, policies, tools, coding and an offline Mind2Web subset. Models choose among supplied admissible actions; they do not generate actions or execute a closed-loop browser policy in the reported evaluation. The paired probes measure selected visible facts, not complete understanding or internal causation. Pooled results are not a web-only leaderboard, and the paper reports coverage limits and label-sensitivity analyses. No benchmark release, code, data or weights were independently inspected.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Pairs factual probes with next-action choices on frozen agent traces, including Mind2Web, to diagnose reading-decision gaps without closed-loop execution.", "topics": ["evaluation", "web-navigation", "interaction-history"], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.18579"}
{"aliases": [], "authors": ["Yuchen Lu", "Julian McAuley", "Lei Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f4a9b6fae506d4dd35798c1930d5df4ec7fc7a2c452ab8bd5a5714cf385d505", "submitted_at": "2026-10-13T17:35:27Z", "updated_at": "2026-10-14T04:46:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.18784v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.18784", "identifiers": {"arxiv": "2610.18784"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.18784"}], "name": "Pack Light, CarryOn: Building Object-Oriented Context for Long-Horizon Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Introduces CarryOn, which maintains evidence-linked task objects and remaining-work context across history compression, with evaluations on desktop tasks and web research.", "topics": ["memory", "long-horizon", "efficiency"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.18784"}
{"aliases": [], "authors": ["Wenbo Hu", "Shraman Pramanick", "Krishna Kumar Singh", "Nanyun Peng", "Kai-Wei Chang", "Yong Jae Lee"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a9f020520d7945cf0034b1701043a806d1550687ed2fc38c55fa7e96cea9377", "submitted_at": "2026-10-13T17:59:51Z", "updated_at": "2026-10-13T17:59:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.18868v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.18868", "identifiers": {"arxiv": "2610.18868"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.18868"}], "name": "Tev: Scaling Long-Horizon Agentic Environments with Dense Progress Rewards", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-13", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.18868"}
{"aliases": [], "authors": ["Qingsong Xie", "Jiahao Tian", "Jiacheng Tang", "Zhifeng Li", "Madalina Ciobanu", "Navan Preet Singh", "Qingqing Mao", "Ritankar Das"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08d90bcec690b238be96207602ca694759c8672fecbed85e30372dd314e4006e", "submitted_at": "2026-10-13T19:37:49Z", "updated_at": "2026-10-13T19:37:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19100v1"}}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.19100", "identifiers": {"arxiv": "2610.19100"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19100"}], "name": "GroundGUI: GUI Grounding by High-Quality Data and Distillation-to-Reinforcement Learning", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces ScreenRef screenshot grounding data and a training sequence from supervised learning through teacher-guided distillation to reinforcement learning for GroundGUI-8B.", "topics": ["grounding", "data-generation", "reinforcement-learning"], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.19100"}
{"aliases": [], "authors": ["Yuxiang Lai", "Jike Zhong", "Chengzhi Liu", "Ming Li", "Haonan Wang", "Xiaofeng Yang", "Liang Zhao", "Youbao Tang", "Ruixi Fan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f72958f9a70e7714f17b0eb51c5453e08ea8ad4e8e7d4caa5e3170231088251", "submitted_at": "2026-10-13T20:01:18Z", "updated_at": "2026-10-13T20:01:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19128v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.19128", "identifiers": {"arxiv": "2610.19128"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19128"}], "name": "CLASHarness: Benchmarking Multi-Agent Conflict in Computer Use via a Controllable and Verifiable Generation Pipeline", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-13", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces CLASHarness, generating verifiable computer-use tasks with conflicting agent reports, shared-state constraints, and coordination requirements to evaluate conflict detection and resolution.", "topics": ["evaluation", "robustness", "workflow"], "updated_at": "2026-10-13", "url": "https://arxiv.org/abs/2610.19128"}
{"aliases": [], "authors": ["Minghao Guo", "Qingyue Jiao", "Xiujin Liu", "Jianhong Pang", "Ruimeng Yang", "Naihao Xu", "Liwei Che", "Haochen Xue", "Ziqi Wang", "Xinying Cai", "Jiaxuan Luo", "Jiaxin Deng", "Xuesong Ye", "Jintao Qu", "Wenyao Gao", "Jingwen Gu", "Chenyang Zhao", "Shichun Liu", "Guibin Zhang", "Zefang Liu", "Siyin Wang", "Wujiang Xu", "Danrui Li", "Zeru Shi", "Woongyeong Yeo", "Shuang Liu", "Wei-Chieh Huang", "Weizhi Zhang", "Tianyu Yang", "Kangsan Kim", "Yebowen Hu", "Bowen Wen", "Bowen Jiang", "Haohao Su", "Ming Yin", "Sheldon Yu", "Mian Zhang", "Yuyan Chen", "Hanrong Zhang", "Kai Mei", "Zihao Zhang", "Mingyu Jin", "Xi Zhu", "Hengyi Wang", "Tianxin Wei", "Linyong Nan", "Zhenwen Liang", "Zhendong Chu", "Ziqi Huang", "Yongye Su", "Sixun Dong", "Shujian Liu", "Kaiqu Liang", "Jiaojiao Han", "Zhaoye Fei", "Kun Wang", "Shihan Dou", "Zhichun Guo", "Xingjian Diao", "Zeqi Gu", "Heng Wang", "Tao Feng", "Cheng Qian", "Hamid Palangi", "Silei Xu", "Xi Yun", "Yi Zhu", "Shilong Liu", "Qingsong Wen", "Sung Ju Hwang", "Zuxuan Wu", "Yu-Gang Jiang", "Yiyu Shi", "Dimitris N. Metaxas", "Jiaxuan You", "Julian McAuley", "Mengdi Wang", "Jiawei Han", "Philip S. Yu", "Ziyi Ye"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f1a48aa4b89fb97e8e1a186ba7345826a8b9efc9aa7b82e0dd1f5ce17d42b73", "submitted_at": "2026-10-14T01:11:57Z", "updated_at": "2026-10-14T01:11:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19366v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.19366v1"], "reason": "Paper-specific document declaration only; no code, data or weights permissions are inferred."}}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T19:02:22.864948+00:00", "role": "primary", "sha256": "308796868bf9b1cb042bc6008e1e6f4ba6e776a2d4e13d40ad5684b477861563", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.19366v1"}], "id": "arxiv:2610.19366", "identifiers": {"arxiv": "2610.19366"}, "kind": "survey", "last_verified_at": "2026-10-20T19:02:22.864948+00:00", "licenses": {"code": null, "data": null, "document": {"basis": "The retained exact-v1 arXiv HTML infobox explicitly declares License: CC BY 4.0 for this document.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T19:02:22.864948+00:00", "url": "https://arxiv.org/html/2610.19366v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19366"}], "name": "Seeing, Maintaining, and Learning: The Evolution of Multimodal Agent Memory", "notes": "A broad literature synthesis spanning images, video, audio, GUI and embodied settings. Its GUI section compares interface-state retention, action outcomes, cross-session reuse and visual evidence, but the paper does not provide a new GUI-agent experiment. Taxonomy corpus counts and benchmark classifications are authors' annotations, not independently audited field coverage. Updating stored content alone does not establish the survey's self-evolving-policy tier; endpoint task success does not isolate memory use. Lifecycle, resource-budget and provenance recommendations are proposed evaluation directions rather than demonstrated guarantees. Platform facets stay unspecified; cited systems and releases were not independently verified.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Surveys multimodal memory using representation and management axes, distinguishing fixed-policy updates from policy evolution and comparing GUI memory evaluation with other modalities.", "topics": ["survey", "memory", "evaluation", "interaction-history"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.19366"}
{"aliases": [], "authors": ["Yuke Wang", "Benlei Cui", "Shen Pang", "Xuemei Dong", "Longtao Huang", "Hui Xue", "Yuwen Zhai", "Junjie Li", "Jingqun Tang", "Haiwen Hong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da334d943d98b4bf9915c6ca28aab734d736f376e67cbab950aa1af65f540ebf", "submitted_at": "2026-10-14T01:30:03Z", "updated_at": "2026-10-14T01:30:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19384v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.19384v1"], "reason": "Paper-specific document declaration only; no code, data or weights permissions are inferred."}}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T19:02:16.329923+00:00", "role": "primary", "sha256": "ca995ec330f62a43d3d4fbbf28dfe83c5fa8b05f5c1e446929493409760268d1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.19384v1"}], "id": "arxiv:2610.19384", "identifiers": {"arxiv": "2610.19384"}, "kind": "paper", "last_verified_at": "2026-10-20T19:02:16.329923+00:00", "licenses": {"code": null, "data": null, "document": {"basis": "The retained exact-v1 arXiv HTML infobox explicitly declares License: CC BY 4.0 for this document.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T19:02:16.329923+00:00", "url": "https://arxiv.org/html/2610.19384v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19384"}], "name": "MMA-SafetyBench: A Benchmark for Multimodal Agent Safety Evaluation", "notes": "ASR measures predicted coordinates inside an annotated target or compromised generated content; actual clicks, environment changes and downstream harm are not verified. SCR additionally counts adopted adversarial intent or task/schema disruption. Clean-task and attacked-trajectory denominators differ, including 42 clean desktop tasks versus 209 adversarial desktop trajectories. Payloads and injection points are fixed, and neutral prompts change both compliance cues and instruction-priority guidance. GPT-5 judging has a 300-trajectory human-label check, with borderline disagreements. Android-style mobile dialogs do not establish an Android runtime. Code/data links were identified but not inspected; no independent reproduction.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Builds trajectory-guided visual-injection tests for web, desktop and mobile observations, separating target-matching action predictions from broader semantic compromise.", "topics": ["evaluation", "safety-security", "grounding"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.19384"}
{"aliases": [], "authors": ["Md Farhan Ishmam", "Kenneth Marino"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d154a14c5ac85fb352721c203fe7bd709eaba4287ef7376e6a6f5fc395708f1", "submitted_at": "2026-10-14T09:06:49Z", "updated_at": "2026-10-14T09:06:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19834v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.19834", "identifiers": {"arxiv": "2610.19834"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19834"}], "name": "Teaching Agents in their Own Style: Blackbox Teacher Distillation for Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces WebKD, which repairs failed student web-agent trajectories with minimal teacher edits and trains on corrections that preserve the student's reasoning style.", "topics": ["data-generation", "web-navigation", "efficiency"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.19834"}
{"aliases": [], "authors": ["Ridong Zhang", "Shuxian Liang", "Chaoqun Wan", "Zhen Huang", "Xian-Sheng Hua"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50d5cb78e6715f0749c5973e900745ecdb3672febc2c8e090da94cb708127352", "submitted_at": "2026-10-14T09:58:40Z", "updated_at": "2026-10-14T09:58:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19900v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:06Z", "role": "primary", "sha256": "652e8369ef11334d76c0eb67afeaa5c9f7841fd31ca02243b0dfb1493316849a", "supports": ["environments", "notes"], "url": "https://iaie-physicalai.github.io/GraceBench"}], "id": "arxiv:2610.19900", "identifiers": {"arxiv": "2610.19900"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19900"}], "name": "How GUI Agents Fail: GraceBench for Execution Stability Diagnosis", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. Android platform tagging is additionally supported by the primary project documentation.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces GraceBench, using application transition graphs and recorded interaction evidence to diagnose deviations, loops, and recovery that final task-success scores can hide.", "topics": ["evaluation", "mobile-navigation", "robustness"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.19900"}
{"aliases": [], "authors": ["Wenkui Yang", "Yiming Hao", "Kai Zhou", "Chengyu Pu", "Yuzhe Zhang", "Jie Cao", "Jin Huang", "Shuai Ma", "Feng Tian"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "602f4eb7aacc05748b83b9c32b28439098005c77eafb7624804f820995521a79", "submitted_at": "2026-10-14T12:35:59Z", "updated_at": "2026-10-14T12:35:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20100v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.20100", "identifiers": {"arxiv": "2610.20100"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20100"}], "name": "Task-State Governance: A Virtual Machine for Verifiable and Recoverable GUI Agent Execution", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents TaskVM, an interface for specifying desired and preserved application state, checking agent actions, and exposing whether changes can be reversed or compensated.", "topics": ["human-in-the-loop", "safety-security", "workflow"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.20100"}
{"aliases": [], "authors": ["Chengguang Gan", "Yunhao Liang", "Hanjun Wei", "Qinghao Zhang", "Zhixi Cai", "Shiwen Ni"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47488d3b5450753b4e1a4b0991e50dc92fbab197a9b54bd8d4e3e1a3d762ee96", "submitted_at": "2026-10-14T12:39:51Z", "updated_at": "2026-10-14T12:39:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20107v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.20107", "identifiers": {"arxiv": "2610.20107"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20107"}], "name": "Why Generate a Click? Plug-in Decision Heads for Web Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Adds decision heads to a frozen vision-language generator for selecting web actions and targets, with a confidence gate that falls back to text generation.", "topics": ["efficiency", "grounding", "web-navigation"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.20107"}
{"aliases": [], "authors": ["Shuo Wang", "Sifan Zhou", "Xiaobo Wang", "Tianyu Fu", "Chenxu Zhao", "Anyang Su", "Hanning Wang", "Ping Jiang", "Minghui Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f9d3ce7dffa8c726ce8129083c8d6c580449d1a32d87c7cd3aaade4fe6cf59b", "submitted_at": "2026-10-14T15:53:48Z", "updated_at": "2026-10-14T15:53:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20390v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.20390", "identifiers": {"arxiv": "2610.20390"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20390"}], "name": "STP-GUI: Spatially-Aware Token Pruning for Efficient Web-GUI Agents", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces STP-GUI, which selects spatially coherent screenshot windows conditioned on the instruction to reduce visual tokens while retaining useful interface structure.", "topics": ["efficiency", "grounding", "screen-understanding"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.20390"}
{"aliases": [], "authors": ["Enhan Zhao", "Yongqi Tong", "Xuejun Yu", "Hang Wang", "Pan Wang", "Yichao Wang", "Di He", "Jianshe Li", "Wei Wu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "def8955ca17baaaa4520d6fcb4b5900406c5dffcfe11b9c4aa7a0c2a0ac5b573", "submitted_at": "2026-10-14T16:16:51Z", "updated_at": "2026-10-14T16:16:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20424v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}], "id": "arxiv:2610.20424", "identifiers": {"arxiv": "2610.20424"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.20424"}], "name": "BizBench: Unveiling Hidden Challenges in Deploying Long-Horizon Agents for Real-World Business", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.20424"}
{"aliases": [], "authors": ["Mingyu Ouyang", "Yiran Wang", "Siyuan Hu", "Xingyilang Yin", "Huiqiang Sun", "Guangzhi Wang", "Junfu Pu", "Cheng Cheng", "Honghao Chen", "Ying Shan", "Mike Zheng Shou", "Hwee Tou Ng"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67b539468bd2bbf0d3b49a5a7eed4791a8ddca4b56ad9515d23a29204d1f069f", "submitted_at": "2026-10-14T17:57:12Z", "updated_at": "2026-10-14T17:57:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20584v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "id": "arxiv:2610.20584", "identifiers": {"arxiv": "2610.20584"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20584"}], "name": "Player-1: Learning to Play from Synthetic Experience", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Extends computer-use policies to game control through synthetic training environments and richer keyboard and pointer actions. Its evaluation concerns games rather than ordinary productivity workflows.", "topics": ["data-generation", "reinforcement-learning", "evaluation"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.20584"}
{"aliases": [], "authors": ["Haofei Yu", "Yixing Chen", "Yining Zhao", "Xuyan Ye", "Yuxiang Liu", "Zhigang Hua", "Yan Xie", "Shuang Yang", "Jiaxuan You"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fede23bfd088bce8edb49f5b7e44b1626f4c5c94999f2ca56ade76fae818db84", "submitted_at": "2026-10-14T17:59:28Z", "updated_at": "2026-10-15T17:52:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.20597v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T04:48:34Z", "role": "primary", "sha256": "4b1219a8d3b34b9ecfabab989eed959ab7e2263500ab2034fc56105db1c559eb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.22673%2C2610.17816%2C2610.20597%2C2610.23857%2C2610.25601&max_results=5"}], "id": "arxiv:2610.20597", "identifiers": {"arxiv": "2610.20597"}, "kind": "paper", "last_verified_at": "2026-10-21T04:48:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.20597"}], "name": "Co-Evolving Acting and Forecasting for Language Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-14", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-15", "url": "https://arxiv.org/abs/2610.20597"}
{"aliases": [], "authors": ["Panrong Tong", "Quyu Kong", "Pengxiang Li", "Tianyu Xia", "Hanzhang Zhou", "Lei Wang", "Chenglin Cai", "Long Li", "Long Chen", "Jianan Zhang", "Zhengwen Zeng", "Xu Zhang", "Gongjie Zhang", "Yue Wang", "Steven Hoi"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5f4eda178718aac211eddc7281d572dc1d189dfd0cd130d4eba9f30f5a05ca2", "submitted_at": "2026-10-14T21:06:01Z", "updated_at": "2026-10-14T21:06:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20939v1"}}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.20939", "identifiers": {"arxiv": "2610.20939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20939"}], "name": "MAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer Environments", "notes": "The annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "Unifies GUI evaluation, trajectory collection, environment synthesis, and online reinforcement learning across mobile, computer, and web tasks.", "topics": ["sandbox", "evaluation", "reinforcement-learning"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.20939"}
{"aliases": [], "authors": ["Hanzhang Zhou", "Long Li", "Quyu Kong", "Lei Wang", "Panrong Tong", "Long Chen", "Tianyu Xia", "Gongjie Zhang", "Xu Zhang", "Chen Liu", "Jianan Zhang", "Chenglin Cai", "Zhengwen Zeng", "Steven Hoi", "Yue Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "beab15a2bf8a6aca19d219dbe0a047e8c6d6bd84a580df33633b5d366c175a0e", "submitted_at": "2026-10-15T16:51:35Z", "updated_at": "2026-10-15T16:51:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22200v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.22200", "identifiers": {"arxiv": "2610.22200"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22200"}], "name": "OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2026-10-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces an open generalist computer-use stack combining visual and programmatic interaction, with model-controlled observation and task-state decisions.", "topics": ["tool-integration", "planning"], "updated_at": "2026-10-15", "url": "https://arxiv.org/abs/2610.22200"}
{"aliases": [], "authors": ["Bowen Yang", "Kaiming Jin", "Zichen Ding", "Zhenyu Wu", "Zehao Li", "Zhoumianze Liu", "Kanzhi Cheng", "Fangzhi Xu", "Yanan Sun", "Qiushi Sun", "Yu Qiao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b7f3542ca2faa63df81e25721a01928bfb35e9e3de458e8be27fd3f9010adc8", "submitted_at": "2026-10-15T17:14:02Z", "updated_at": "2026-10-15T17:14:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22234v1"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.22234", "identifiers": {"arxiv": "2610.22234"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22234"}], "name": "OS-Helix: Unifying GUI Interaction and Code Execution in Native Computer-Using Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2026-10-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Trains native computer-use models to coordinate GUI actions and code execution using hybrid demonstrations and online reinforcement learning.", "topics": ["tool-integration", "reinforcement-learning"], "updated_at": "2026-10-15", "url": "https://arxiv.org/abs/2610.22234"}
{"aliases": [], "authors": ["Qingni Wang", "Changdae Oh", "Yuzhe Yang", "Sharon Li", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a249a56a61ebb14b456500f9b245efe7b72997af5d5f515e5cc7707fe68b018", "submitted_at": "2026-10-15T17:40:12Z", "updated_at": "2026-10-15T17:40:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22280v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.22280", "identifiers": {"arxiv": "2610.22280"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22280"}], "name": "LEASH: Uncertainty-Aware Weak-to-Strong Escalation and Oversight for Computer-Use Agents", "notes": "Editorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.", "published_at": "2026-10-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Uses weak-agent uncertainty to decide when to escalate to a stronger model and monitor its subsequent task execution.", "topics": ["efficiency", "planning"], "updated_at": "2026-10-15", "url": "https://arxiv.org/abs/2610.22280"}
{"aliases": [], "authors": ["Tengda Guo", "Jiawang Hu", "Yifei Tao", "Yin Wang", "Haotian Hu", "Fanyi Wang", "Liujuan Tang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "025c7018c86ae4de79a659242070b2f91c5913333e9ecd4bd7fd357d13d27a34", "submitted_at": "2026-10-14T14:39:41Z", "updated_at": "2026-10-14T14:39:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22408v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T12:58:20.217517Z", "role": "primary", "sha256": "f275c47a8b6d349a70e39a9e55eea5a09d99d5f35cc896100f3bcfd5b0923e0b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202609200000+TO+202610201255%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:54.814500Z", "role": "primary", "sha256": "0389d96a8f6c4975129f5c526e8df9ca4690f31364a2f4e66aaea4a75f0eabbc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.22408v1"}], "id": "arxiv:2610.22408", "identifiers": {"arxiv": "2610.22408"}, "kind": "paper", "last_verified_at": "2026-10-20T13:12:54.814500Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22408"}], "name": "ECHOES: Seeing Beyond the Current Image via Recursive History Compression and Privileged On-Policy Distillation", "notes": "Privileged screenshot history is used during training, while deployment retains self-generated textual state. GUI results use action/step metrics and model-specific history settings; online Sokoban transfer is separate. GRPO rewards action correctness, valid output format and compression length. Diagnostic memory replay is distinct from the validation gap used to schedule training. No independent reproduction.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "core", "summary": "ECHOES learns recursive textual task memory through privileged screenshot-history distillation and GRPO, then predicts GUI actions from the current screenshot and memory.", "topics": ["memory", "interaction-history", "state-tracking", "long-horizon", "imitation-learning", "reinforcement-learning", "efficiency", "evaluation"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.22408"}
{"aliases": [], "authors": ["Ruida Wang", "Gerald Friedland", "Tong Zhang", "Chris Kong"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "675d07a096cac3a71fde043d0c613ea212a8e8e6f5f4172d92d2db08579f9ac3", "submitted_at": "2026-10-15T15:26:54Z", "updated_at": "2026-10-15T15:26:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22448v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-19T21:04:13Z", "role": "primary", "sha256": "195764703dc88bfa85d038661598fac5c25b0019d771435cc33efc38841751e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.22448&max_results=1"}], "id": "arxiv:2610.22448", "identifiers": {"arxiv": "2610.22448"}, "kind": "paper", "last_verified_at": "2026-10-19T21:04:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.22448"}], "name": "Lean4AgenticRL: Advancing Agentic RL with Verifiable Multi-level Process Rewards", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-15", "url": "https://arxiv.org/abs/2610.22448"}
{"aliases": [], "authors": ["Md Farhan Ishmam", "Minh Pham Dinh", "Kenneth Marino"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f064032a0b0d673273d90a9250c76a6b823b1728761635b8a555a0bd96b888a", "submitted_at": "2026-10-15T18:37:41Z", "updated_at": "2026-10-15T18:37:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22539v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-19T04:42:05Z", "role": "primary", "sha256": "ffbb3f28a3a0fc8888f30ea8b2cdaf441239685aaacc2d80f311581f4b5a1d5c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610050000+TO+202610190000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T10:34:07.486792+00:00", "role": "primary", "sha256": "0c95ddb42bbfc91aabf25e2dc2d2393a9977b6d979a2c891c5bcf80798dbfbc3", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.22539v1"}], "id": "arxiv:2610.22539", "identifiers": {"arxiv": "2610.22539"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22539"}], "name": "Don't Retrain, Remix: Adaptation of Web Agents using Mixture Hypernetworks", "notes": "Reviewed the complete retained abstract and selected exact-v1 method, evaluation and analysis sections. Adapter-bank and router training use source demonstrations. Domain scouting accesses target training-task intents and answer-stripped reference procedures; version scouting is task-free. No target-time gradients does not mean no target task information. No independent reproduction or artifact/license verification.", "published_at": "2026-10-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "WebMix uses a scout-produced interface manual to route mixtures of source-trained LoRA adapters for web agents, adapting to target interfaces without target-time gradient updates.", "topics": ["web-navigation", "imitation-learning", "robustness"], "updated_at": "2026-10-15", "url": "https://arxiv.org/abs/2610.22539"}
{"aliases": [], "authors": ["Adamo Orsini", "Seyed Mehdi Ayyoubzadeh", "Sadman Rakib Pinon", "Hannuo Zhang", "Koosha Gholipour Baradari", "Yuanhao Yu", "Xinxin Zuo", "Yang Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "caaf24ef4d8b86c8aaebc852235d0fe947a5ed859dc1ba8329ce07a2f7be9aa1", "submitted_at": "2026-10-15T21:40:34Z", "updated_at": "2026-10-15T21:40:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22672v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:56:56.464343+00:00", "role": "primary", "sha256": "5fd9d7926e99d068fa8fb7238a9e7811e3f8c55d00e6b89d5672754cb3e619bc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.22672v1"}], "id": "arxiv:2610.22672", "identifiers": {"arxiv": "2610.22672"}, "kind": "paper", "last_verified_at": "2026-10-20T12:56:56.464343+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22672"}], "name": "Not All Graphs Are Maps: A Task-Agnostic Exploration Framework for Mobile GUI Navigation", "notes": "The controlled cohort has 84 tasks across six apps; the broader inventory spans 20 apps. Main results use human destination labels. DroidTask and SPA-Bench are converted to navigation-only protocols. Some comparison cells received another graph condition’s memory, with sensitivity analyses reported. Cross-family runs differ in app versions, state and denominators. A small stale-graph probe does not establish robustness. Models, code, artifacts and performance were not independently reproduced.", "published_at": "2026-10-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "TAG-Nav supplies Android GUI agents with destination and route memory distilled from offline app exploration. DestBench measures arrival at a reference pre-action screen, separately from downstream task execution.", "topics": ["mobile-navigation", "memory", "evaluation"], "updated_at": "2026-10-15", "url": "https://arxiv.org/abs/2610.22672"}
{"aliases": [], "authors": ["Chloe Su", "Jiahan Li", "Zhenting Qi", "Chaoran Cheng", "Haoyu Chen", "Sejoon Jun", "Weirui Ye", "Yilun Du"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0d88e9187454ddb609b6de5345bd03b961a29df22e220c6673fd55b3dda9051", "submitted_at": "2026-10-15T21:56:00Z", "updated_at": "2026-10-15T21:56:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22686v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-22T04:46:45Z", "role": "primary", "sha256": "aaabbb5ccd1734a9eb08f60f041c82b3d0f125105e9ebe30fe554724466c27a3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.22686%2C2610.26679&max_results=2"}], "id": "arxiv:2610.22686", "identifiers": {"arxiv": "2610.22686"}, "kind": "paper", "last_verified_at": "2026-10-22T04:46:45Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.22686"}], "name": "Multimodal Filesystem Memory For Long-Horizon Embodied Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-15", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-15", "url": "https://arxiv.org/abs/2610.22686"}
{"aliases": [], "authors": ["Umar Iqbal"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "077dd623c980813e83197a38b6545161a3d2fc842d99fe291c53372b9770acd2", "submitted_at": "2026-10-15T22:02:45Z", "updated_at": "2026-10-15T22:02:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22697v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:38:46.444001Z", "role": "primary", "sha256": "cb6a8a9c0cc47be21b36b3d1831b26c6189d8f2e4af0d7e26d643c9c2a65452f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.22697v1"}], "id": "arxiv:2610.22697", "identifiers": {"arxiv": "2610.22697"}, "kind": "paper", "last_verified_at": "2026-10-20T13:38:46.444001Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22697"}], "name": "What It Is, Not What It Says: Securing Agentic Browsing Interactions with Web Semantics", "notes": "Evaluation instructs the model to simulate compromise and measures planner outputs and permission appearance. Utility uses plan matching and LLM-judged element suitability; no steps are executed in that assessment. Assumes an honest first party, trusted browser/harness and untampered instrumentation; multi-page and availability attacks are outside scope. Parallel latency reduction is an unimplemented estimate. No independent reproduction.", "published_at": "2026-10-15", "relations": [], "review_status": "curated", "scope": "core", "summary": "Intact records webpage element provenance and capabilities, isolates planning per element, and uses a browser harness to constrain actions and request consent for third-party interactions.", "topics": ["safety-security", "web-navigation", "planning", "human-in-the-loop"], "updated_at": "2026-10-15", "url": "https://arxiv.org/abs/2610.22697"}
{"aliases": [], "authors": ["Haoming Li", "Defu Cao", "Xu Hu", "Yingtao Luo", "Yiyang Wang", "Xianghe Xu", "Jiashu He", "Wanting Jiang", "Yifan Zhang", "Yunhan Zhao", "Siyang Wu", "Chen Yang", "Haixin Wang", "Zheyuan Deng", "Taoran Li", "Jingwen Zhang", "Zhejian Peng", "Shushu Liang", "Yue Zhao", "Bingzhe Li", "Feng Chen", "Yifan Wang", "Jintao Huang", "Yijun Wang", "Yuexing Hao", "Xiaomin Li"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f613ef7ada260b494a411058bc733c7a5e0b074cdb01906c38132acf7275b9b1", "submitted_at": "2026-10-16T06:14:55Z", "updated_at": "2026-10-16T06:14:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22990v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-19T04:42:05Z", "role": "primary", "sha256": "ffbb3f28a3a0fc8888f30ea8b2cdaf441239685aaacc2d80f311581f4b5a1d5c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610050000+TO+202610190000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T10:42:06.870916+00:00", "role": "primary", "sha256": "bf9066e8b3472100f120e21534f92655f4d0814f2521b9d23d778e5239c7a2a4", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.22990v1"}], "id": "arxiv:2610.22990", "identifiers": {"arxiv": "2610.22990"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22990"}], "name": "PersonaCUABench: Benchmarking Computer-Use Agents Serving Diverse Personas", "notes": "Reviewed the complete retained abstract and selected exact-v1 construction, evaluation and limitations sections. Main interactions and satisfaction ratings use LLM-simulated personas; human annotators validate selected outputs. The harness observes page text, with screenshots only for auditing. Results are not evidence about real demographic populations or desktop agents. No independent reproduction.", "published_at": "2026-10-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "PersonaCUABench varies simulated user requests and clarification dialogue over fixed web tasks, comparing task completion with requirement handling and persona-agent satisfaction ratings.", "topics": ["evaluation", "web-navigation", "human-in-the-loop"], "updated_at": "2026-10-16", "url": "https://arxiv.org/abs/2610.22990"}
{"aliases": [], "authors": ["Chulhoon Yoon", "Kyungri Park", "Jinyeong Yun", "Daeyun Jeong", "Sukmin Yun", "Woohwan Jung"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9585696d256bacc710dce57bfd7cda6a8037ce5725a30f959028352bc07e97c", "submitted_at": "2026-10-16T06:44:56Z", "updated_at": "2026-10-16T06:44:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23012v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-19T04:42:05Z", "role": "primary", "sha256": "ffbb3f28a3a0fc8888f30ea8b2cdaf441239685aaacc2d80f311581f4b5a1d5c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610050000+TO+202610190000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T10:34:04.380045+00:00", "role": "primary", "sha256": "75c025a862a046b8bcbdd84b800a37132edcc1ec271d40c57c18ec984652a0b0", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.23012v1"}], "id": "arxiv:2610.23012", "identifiers": {"arxiv": "2610.23012"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.23012"}], "name": "Adaptation of Computer Use Agents to New Software with User Manuals", "notes": "Reviewed the complete retained abstract and selected exact-v1 methods, evaluation and validation sections. Execution plans stay fixed per episode; policy training is separate and application-specific. Screenshot-based training/Slack judgments can miss task conditions, and OSWorld tasks may overlap backbone pretraining. No independent reproduction, safe-deployment guarantee or artifact-release verification.", "published_at": "2026-10-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "MAGE retrieves software-manual sections into fixed execution plans and generates setup/main task-plan pairs whose GUI rollouts support application-specific policy adaptation with round-based GRPO.", "topics": ["planning", "data-generation", "reinforcement-learning"], "updated_at": "2026-10-16", "url": "https://arxiv.org/abs/2610.23012"}
{"aliases": [], "authors": ["Songping Wang", "Borui Li", "Qinglong Liu", "Yueming Lyu", "Yi Huang", "Fengting Li", "Chao Feng", "Caifeng Shan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea753c1cb0666950d89f2781511f26872e3d5b1c2fd2434896248214160fa22c", "submitted_at": "2026-10-16T11:06:37Z", "updated_at": "2026-10-16T11:06:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23303v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T10:42:09.967883+00:00", "role": "primary", "sha256": "ec9d647818be346ddbf7807be87b4789cd53e50c5d3d7a7904ad4f49bc71c38e", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.23303v1"}], "id": "arxiv:2610.23303", "identifiers": {"arxiv": "2610.23303"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.23303"}], "name": "MobileJailBench: Red Queen Jailbreak Benchmarking for Mobile Agents", "notes": "Reviewed the complete retained abstract and selected exact-v1 threat-model, construction and evaluation sections. Execution uses a browser-based Android-like simulator, not real Android isolation or live services. The attacker controls instructions and designated carriers, not privileged simulator state/APIs or system/model/harness internals. Rates depend on release, budget and adjudication coverage. SAF labels track observable expressions and actions, not latent reasoning; unresolved episodes are excluded from rate denominators. No independent reproduction.", "published_at": "2026-10-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileJailBench evaluates mobile-agent jailbreaks in a browser-based Android-like simulator, combining feedback-driven attack evolution, separately admitted task expansions and evidence-grounded outcome adjudication.", "topics": ["safety-security", "mobile-navigation", "evaluation", "observability", "cross-app"], "updated_at": "2026-10-16", "url": "https://arxiv.org/abs/2610.23303"}
{"aliases": [], "authors": ["Songping Wang", "Peiyang Li", "Yi Huang", "Fengting Li", "Xuhao Ran", "Yueming Lyu", "Chao Feng", "Caifeng Shan"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3fc22fbeda19ac2947b00af73801f0674c550e0d7c78919f811cae754b06128c", "submitted_at": "2026-10-16T11:10:30Z", "updated_at": "2026-10-16T11:10:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23307v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:36.223826+00:00", "role": "primary", "sha256": "3803984a185b6258276b87b87ac1991583878fb41dc22618ff743f6f69c7bc7e", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.23307v1"}], "id": "arxiv:2610.23307", "identifiers": {"arxiv": "2610.23307"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.23307"}], "name": "MobileInject: Evaluating Reasoning-to-Action Security of MLLM-Based Mobile Agents under Cross-Modal Consistency Injection", "notes": "Reviewed the complete abstract and selected v1 construction, threat-model and evaluation sections. The attacker modifies observation surfaces, not system instructions, model parameters or the OS. State oracles are distinct from emitted-rationale diagnoses, whose labels include outcome-dependent constraints. Those diagnostics are not independent evidence of hidden-reasoning causality. No reproduction or release verification.", "published_at": "2026-10-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "MobileInject evaluates visual/textual observation injection in stateful Android workflows using resettable emulators, benign/adversarial state oracles and trace-based diagnosis of emitted rationales and actions.", "topics": ["safety-security", "mobile-navigation", "evaluation", "observability", "cross-app"], "updated_at": "2026-10-16", "url": "https://arxiv.org/abs/2610.23307"}
{"aliases": [], "authors": ["Qianli Ma", "Limengxi Yue", "Shuyang Xu", "Yanpei Shi", "Hongrui Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8922fb54b82e6825cb3a841f2400c7773354a8a49db5e31035dc468f3bdf96b3", "submitted_at": "2026-01-12T01:04:20Z", "updated_at": "2026-01-12T01:04:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23857v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T04:48:34Z", "role": "primary", "sha256": "4b1219a8d3b34b9ecfabab989eed959ab7e2263500ab2034fc56105db1c559eb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.22673%2C2610.17816%2C2610.20597%2C2610.23857%2C2610.25601&max_results=5"}], "id": "arxiv:2610.23857", "identifiers": {"arxiv": "2610.23857"}, "kind": "paper", "last_verified_at": "2026-10-21T04:48:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.23857"}], "name": "Web Agent Agentic Reinforcement Learning Decision Model Under Multi-Cost and Failure Risk", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-01-12", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-01-12", "url": "https://arxiv.org/abs/2610.23857"}
{"aliases": [], "authors": ["Xinyu Fang", "Yuchi Wang", "Weihao Yu", "Rongyao Fang", "Zhaohai Li", "Zhixun Li", "Tiancheng Gu", "Xilin Wei", "Yuhang Zang", "Hongsheng Li", "Dahua Lin", "Guofeng Zhang", "Shuai Bai", "Zhibo Yang", "Dayiheng Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6dfa0c59ec1262cebcf2a4e61ab1e0b58c647513d8e6eedb2ce7e6a1a675558c", "submitted_at": "2026-10-14T12:46:54Z", "updated_at": "2026-10-14T12:46:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23986v1"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T13:48:11Z", "role": "primary", "sha256": "89e110b6fa882639fb73936cb27e5373d11b632938c31968f563152534bf5670", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.14966%2C2610.23986&max_results=2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:09:58.148629Z", "role": "primary", "sha256": "bef4399f087e484b62ff5b8dc2c283f92e0357b94c93b464b86fd97fe7b069cb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.23986v1"}], "id": "arxiv:2610.23986", "identifiers": {"arxiv": "2610.23986"}, "kind": "paper", "last_verified_at": "2026-10-20T15:09:58.148629Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.23986"}], "name": "BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References", "notes": "The evaluated agents use Linux shell/Python/MCP tools and can inspect viewport screenshots; graphical CUA control and Windows/macOS are future work. Each agent is run once per task, with different Codex/Claude Code harnesses across systems. Scores combine weighted rubric satisfaction and model-rated visual quality. Tool counts include unsuccessful calls and do not establish matched compute or successful actions. No independent reproduction or artifact-release verification.", "published_at": "2026-10-14", "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Benchmarks agents creating editable Blender projects through shell, Python and MCP tools, judging delivered scenes with multi-view renders, task-specific rubrics and visual-quality ratings.", "topics": ["evaluation", "tool-integration", "sandbox"], "updated_at": "2026-10-14", "url": "https://arxiv.org/abs/2610.23986"}
{"aliases": [], "authors": ["Egor Zverev", "Juhee Kim", "Leon Trampert", "Boris Köpf", "Aashish Kolluri", "Shruti Tople", "Santiago Zanella-Béguelin"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e675a3082f571da547fb5b6cb89e4f68a034bfa57236886e1218cb2bb2cdf1c3", "submitted_at": "2026-10-16T22:12:45Z", "updated_at": "2026-10-16T22:12:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.24277v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:33.132448+00:00", "role": "primary", "sha256": "daceefecb2e6b910a6c7f2ffe35e89c7070f6282f4f35ee89e885a2e77aa94c4", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.24277v1"}], "id": "arxiv:2610.24277", "identifiers": {"arxiv": "2610.24277"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.24277"}], "name": "SPIDES: Information-Flow Control for Browser Agents", "notes": "Reviewed the complete abstract and selected v1 threat-model, policy, evaluation and limitations sections. Guarantees assume correct labels/trusted origins and a trusted browser/harness mediating actions; evaluation uses idealized approvals. Origin-level reader sets can miss recipient distinctions within one origin. No unrestricted prompt-injection immunity or independent reproduction is claimed.", "published_at": "2026-10-16", "relations": [], "review_status": "curated", "scope": "core", "summary": "SPIDES propagates HTML integrity and confidentiality labels into agent context and mediates browser actions with information-flow policies and approval gates, evaluated in the labeled MicroNet ecosystem.", "topics": ["safety-security", "web-navigation", "human-in-the-loop", "evaluation"], "updated_at": "2026-10-16", "url": "https://arxiv.org/abs/2610.24277"}
{"aliases": [], "authors": ["Sizhe Tang", "Hanhan Zhou", "Jeffrey Svajlenko", "Rashmi Gangadharaiah"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9d6808cf26a609b50b7835ae3ecece0a13580aafcd589823f0b1eeff54732f69", "submitted_at": "2026-10-17T02:03:20Z", "updated_at": "2026-10-17T02:03:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.24372v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T12:58:15.922053Z", "role": "primary", "sha256": "b379e14eefa272f62641224fe1d839b7dfbc414d957124b8abeb103444a843ef", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3AGUI+OR+ti%3A%22computer-using%22+OR+ti%3A%22agentic+browsing%22+OR+ti%3A%22visual+grounding%22+OR+ti%3A%22UI+automation%22%29+AND+submittedDate%3A%5B202608210000+TO+202610201255%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:51.579373Z", "role": "primary", "sha256": "9dff18935a1faa7ec16366bac6a6e9e41f0381d561568c3aba39fe783efcfbd3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.24372v1"}], "id": "arxiv:2610.24372", "identifiers": {"arxiv": "2610.24372"}, "kind": "paper", "last_verified_at": "2026-10-20T13:12:51.579373Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.24372"}], "name": "Beyond Next-Screen Prediction: The Decision Value of GUI World Models", "notes": "The matched augmentation experiment measures offline ranking; AndroidWorld task-success comparisons use differently trained evaluators and do not isolate prediction value. A separate web evaluator is trained on Mind2Web, and web results do not uniformly favor direct scoring. The main study concerns one-step selection. No independent reproduction.", "published_at": "2026-10-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Separates next-screen prediction from action evaluation in GUI lookahead and trains Direct Action Ranking on reference actions, comparing prediction-augmented scoring under matched offline supervision.", "topics": ["evaluation", "world-models", "action-prediction", "imitation-learning"], "updated_at": "2026-10-17", "url": "https://arxiv.org/abs/2610.24372"}
{"aliases": [], "authors": ["Yutong Song", "Jiang Wu", "Yuxuan Ren", "Tony Qin", "Baoyuan Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c33c03c3ea483909eee134d737a77af59ca1fcc089f2eff030a3fff910e9d2a", "submitted_at": "2026-10-17T04:43:26Z", "updated_at": "2026-10-17T04:43:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.24453v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T19:13:17Z", "role": "primary", "sha256": "0ae4417a09822b1d907ff195c656f8f1419ee90d781ec4e80436b990875358de", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22+OR+abs%3A%22web+navigation%22+OR+abs%3AAndroidWorld+OR+abs%3AMind2Web%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.24453", "identifiers": {"arxiv": "2610.24453"}, "kind": "paper", "last_verified_at": "2026-10-20T19:13:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.24453"}], "name": "Hippo: Training-Free Agent Self-Evolution via Hippocampal Replay", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-17", "url": "https://arxiv.org/abs/2610.24453"}
{"aliases": [], "authors": ["Sanket Mhatre", "Sumit Gulwani", "Vu Le", "Yasharth Bajpai", "Gust Verbruggen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6581c87a59b65b764870cf98496512e9687f2ded254904dc6cdd021fd9cfe10a", "submitted_at": "2026-10-17T14:49:08Z", "updated_at": "2026-10-17T14:49:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.24729v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}], "id": "arxiv:2610.24729", "identifiers": {"arxiv": "2610.24729"}, "kind": "paper", "last_verified_at": "2026-10-20T05:52:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.24729"}], "name": "fxBench: Evaluating and Understanding Formula Suggestions in Spreadsheets", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-17", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-17", "url": "https://arxiv.org/abs/2610.24729"}
{"aliases": [], "authors": ["Hridya Dhulipala", "Xiaokai Rong", "Aashish Yadavally", "Tien N. Nguyen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2e03ebe24ed7c3d3283d4c71f9244bf42e682f4e2ffa4f8e4482cfcab67b8ddb", "submitted_at": "2026-10-19T00:07:30Z", "updated_at": "2026-10-19T00:07:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.25601v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T04:48:34Z", "role": "primary", "sha256": "4b1219a8d3b34b9ecfabab989eed959ab7e2263500ab2034fc56105db1c559eb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.22673%2C2610.17816%2C2610.20597%2C2610.23857%2C2610.25601&max_results=5"}], "id": "arxiv:2610.25601", "identifiers": {"arxiv": "2610.25601"}, "kind": "paper", "last_verified_at": "2026-10-21T04:48:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.25601"}], "name": "Step-Aligned Script Generation with Verification and Recovery for Web Automation Testing", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.25601"}
{"aliases": [], "authors": ["Zhanhui Zhou", "Weichen Zhang", "Haoran Liu", "Lingjie Chen", "Tianneng Shi", "Kevin Lin", "Zhengyuan Yang", "Lijuan Wang", "Dawn Song"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a297401f7ad4dcd2f485ade6bab295e365280bb803dcecc111aebf0693dd203", "submitted_at": "2026-10-19T09:30:54Z", "updated_at": "2026-10-19T09:30:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26071v1"}}, "environments": ["desktop", "linux", "windows", "web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:28:55.967592+00:00", "role": "primary", "sha256": "d4231fa9677e916ca8cc27493dfbbe9fb34c7430207150e6b9be10c2bad43bfb", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26071v1"}], "id": "arxiv:2610.26071", "identifiers": {"arxiv": "2610.26071"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26071"}], "name": "CUA-Lite: An Open Platform for Computer Use Agents", "notes": "Reviewed the complete official abstract and selected v1 methods, training and limitations sections. WebVoyager uses offline mirrors; VM-free sandboxes are Linux-only, and reported desktop/mobile gains do not establish unseen-app transfer. No results or released artifacts were independently reproduced.", "published_at": "2026-10-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "CUA-Lite unifies environment interfaces, supervised data formats and per-model harnesses for evaluation, reinforcement learning and supervised fine-tuning across desktop, browser and mobile tasks.", "topics": ["tool-integration", "data-generation", "evaluation", "imitation-learning", "reinforcement-learning", "sandbox"], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.26071"}
{"aliases": [], "authors": ["Fuyao Zhang", "Wenwen He", "Jiaming Zhang", "Yurong Hao", "Xuan Wang", "Guowei Guan", "Che Wang", "Longtao Huang", "Wei Yang Bryan Lim"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "41178d19a23653c0f49423b8f725584bded7ab8819a1a46e83a062794460b4cb", "submitted_at": "2026-10-19T11:30:48Z", "updated_at": "2026-10-19T11:30:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26203v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:23.872790+00:00", "role": "primary", "sha256": "28848f8f597b2a8865ad98f5aa1e4401e15a49f92b33f7bce4d4f044c168ada0", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26203v1"}], "id": "arxiv:2610.26203", "identifiers": {"arxiv": "2610.26203"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26203"}], "name": "Focused Omni-Modal GUI Agents", "notes": "Reviewed the complete retained abstract and selected v1 method/evaluation sections. OmniGUI uses episode-disjoint reference-action evaluation; Type Match, Exact Match and Goal Progress are not presented as independently demonstrated live task-success gains. The Android role is explicit in Appendix D.1. No reproduction or artifact-release verification.", "published_at": "2026-10-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "FocusGUI learns step-specific weights over screenshot, audio, video and interaction-context channels, injecting a compact residual into a frozen decoder without pruning its inputs.", "topics": ["screen-understanding", "grounding", "imitation-learning", "interaction-history", "mobile-navigation"], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.26203"}
{"aliases": [], "authors": ["Zhizhou Wang", "Yaxi Lu", "Zhong Zhang", "Zikun Hu", "Xin Cong", "Yankai Lin", "Fangming Liu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08ef7c9cdda652f60d3f2d4ab89e7218ad5a2b85f5c76f0277c289db06fd2796", "submitted_at": "2026-10-19T14:17:33Z", "updated_at": "2026-10-19T14:17:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26469v1"}}, "environments": ["desktop", "linux", "web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:28:59.073204+00:00", "role": "primary", "sha256": "f13f0225c721d5aa430bb8aea1d1e7a7da5973760c4c3b82f26ba648aade82ba", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26469v1"}], "id": "arxiv:2610.26469", "identifiers": {"arxiv": "2610.26469"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26469"}], "name": "VeriCUA: Learning Proactive Verification for Reliable GUI Agent Rewards", "notes": "Reviewed the complete abstract and selected v1 method/evaluation sections. The Judge is trained on OSWorld and frozen for MobileWorld/WebArena transfer; downstream Actor RL is separate. Evaluation uses screened terminal outcomes, and reported cost counts Probe tokens. No independent reproduction.", "published_at": "2026-10-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "VeriCUA trains a Judge to select verification checks and stopping decisions while a frozen Probe inspects the post-task GUI. Single-attempt reinforcement learning balances verdict correctness and Probe use.", "topics": ["reward-models", "reinforcement-learning", "evaluation", "efficiency"], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.26469"}
{"aliases": [], "authors": ["Zhicheng Zheng", "Cheng Zhang", "Shengping Yu", "Tianxiang Zhao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7dc2ceb9aadbf7506efc068df99326fea5b60fdc6feb340c490841b88aa19c30", "submitted_at": "2026-10-19T14:38:41Z", "updated_at": "2026-10-19T14:38:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26503v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:29:02.225995+00:00", "role": "primary", "sha256": "0554aa531a5f94f51cba6c51bd1b13fb05b0f462497afa5d1dae7ec6fc6e375a", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26503v1"}], "id": "arxiv:2610.26503", "identifiers": {"arxiv": "2610.26503"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26503"}], "name": "ScopeMem: Latent Memory That Learns Where to Revise Frozen GUI Agents", "notes": "Reviewed the complete abstract and selected v1 controlled-study, method and limitations sections. Results are offline, step-level action matching; online task success is explicitly left untested. The frozen actor is queried twice per step. No independent reproduction or redistribution-license claim.", "published_at": "2026-10-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "ScopeMem conditions latent-memory prefixes on a frozen GUI agent's initial action and trains them to correct errors while preserving correct decisions. Its evaluation separates action repairs from harmful revisions.", "topics": ["memory", "interaction-history", "imitation-learning", "robustness", "evaluation"], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.26503"}
{"aliases": [], "authors": ["Zhuomin Chen", "Jingchao Ni", "Xu Zheng", "Yong Chen", "Mo Sha", "Wei Cheng", "Dongsheng Luo"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e91c0f0ab28f7f2da89bdff5f9ce3dc592469a8606f35137144f514e70c96f78", "submitted_at": "2026-10-19T16:18:16Z", "updated_at": "2026-10-19T16:18:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26660v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:26.960480+00:00", "role": "primary", "sha256": "7cb553f8230cfb203df446f08e29d48b676addfe9b076d35171d09970d1970bf", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26660v1"}], "id": "arxiv:2610.26660", "identifiers": {"arxiv": "2610.26660"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26660"}], "name": "Tandem: Serving Hybrid Linear-Attention GUI Agents on Mobile NPUs", "notes": "Reviewed the complete abstract and selected v1 system/evaluation sections. Main latency measurements replay recorded requests and exclude app execution; the paper also reports instrumented real tasks. These are serving-time, energy and action-fidelity results, not a claimed task-success improvement. Hardware, cooling and model configurations matter. No independent reproduction.", "published_at": "2026-10-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "Tandem serves hybrid GUI-agent models on Android NPUs using divergence-point state checkpoints, rollback-aware speculative decoding and idle power release between request bursts.", "topics": ["efficiency", "mobile-navigation"], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.26660"}
{"aliases": [], "authors": ["Hossein A. Rahmani", "Bin Wu", "Chaeeun Lee", "Fanghua Ye", "Sahan Bulathwela", "Sujay Kumar Jauhar", "Marzieh Fadaee", "Nick Craswell", "Emine Yilmaz"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ea5f7386b9cc04494875171fa25b1374922313c37ed5f01341e6d515c984859", "submitted_at": "2026-10-19T16:27:14Z", "updated_at": "2026-10-19T16:27:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26679v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-22T04:46:45Z", "role": "primary", "sha256": "aaabbb5ccd1734a9eb08f60f041c82b3d0f125105e9ebe30fe554724466c27a3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.22686%2C2610.26679&max_results=2"}], "id": "arxiv:2610.26679", "identifiers": {"arxiv": "2610.26679"}, "kind": "paper", "last_verified_at": "2026-10-22T04:46:45Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.26679"}], "name": "StepBench: Diagnosing Real-World Procedural Understanding in AI Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.26679"}
{"aliases": [], "authors": ["Ananya Gubbi Mohanbabu", "Rosiana Natalie", "Anhong Guo", "Amy Pavel"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9858b5ec15ee953504bd00e26b455c1e81104f80429c2dd112de17f9d23977ab", "submitted_at": "2026-10-19T16:39:11Z", "updated_at": "2026-10-19T16:39:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26703v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:30.049627+00:00", "role": "primary", "sha256": "9a964cbf50123887ad20bf156e69c0bbe4567d6b2c28e4cf202d5538eab8294c", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26703v1"}], "id": "arxiv:2610.26703", "identifiers": {"arxiv": "2610.26703"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26703"}], "name": "Comparing Computer Use Behavior Across Multiple Levels of Abstraction", "notes": "Reviewed the complete abstract and selected v1 taxonomy, annotation, evaluation and limitations sections. MAP analyzes existing human/agent traces rather than deploying a new acting policy. Annotation and timing coverage depend on the source datasets; manual agreement reflects a particular labeling scheme. Behavioral findings do not establish a privacy guarantee. No independent reproduction.", "published_at": "2026-10-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "MAP annotates computer-use traces at linked operation, function/consequence, method and goal levels, then compares human and agent workflows through sequence, pattern and effort metrics.", "topics": ["evaluation", "observability"], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.26703"}
{"aliases": [], "authors": ["Zhaoyang Wang", "Qianhui Wu", "Haorui Wang", "Simon Yu", "Baolin Peng", "Canwen Xu", "Hao Cheng", "Wenlin Yao", "Boyi Liu", "Yite Wang", "Jianfeng Gao", "Huaxiu Yao"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f1ff33859352b61f017709fff91fd8e32f08636e9423fb044e7ca8e28f06905", "submitted_at": "2026-10-19T17:55:35Z", "updated_at": "2026-10-19T17:55:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26828v1"}}, "environments": ["web", "desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:29:05.337543+00:00", "role": "primary", "sha256": "b680e07a5a883e7bb0210a1734a628be6c47d91b7df0487a2044d56d1b7e227e", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26828v1"}], "id": "arxiv:2610.26828", "identifiers": {"arxiv": "2610.26828"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26828"}], "name": "Agent World Model in the Wild: Computer-Use Agent Environment Synthesis", "notes": "Reviewed the complete abstract and selected v1 synthesis, training and evaluation sections. The system uses hybrid GUI/MCP/CLI interfaces. Evaluation includes benchmark subsets and model-based judgments, so reported scores are not treated as directly comparable GUI-only results. No independent reproduction.", "published_at": "2026-10-19", "relations": [], "review_status": "curated", "scope": "core", "summary": "WildAWM synthesizes resettable GUI/MCP/CLI environments and cross-interface tasks, using execution-based checks of solvability, difficulty and verifier consistency before supervised and reinforcement learning.", "topics": ["data-generation", "sandbox", "tool-integration", "long-horizon", "imitation-learning", "reinforcement-learning"], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.26828"}
{"aliases": [], "authors": ["Ocean Monjur", "Daud Ahmad Nisar", "Shahriar Kabir Nahin", "Tadesse Destaw Belay", "Sushruti Mishra", "Israel Abebe Azime", "Anshuman Chhabra"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15531788fe6a9dab5a008ffc1d660fc2680bcf132215f1ce874ab36af9bb9a7a", "submitted_at": "2026-10-19T19:40:15Z", "updated_at": "2026-10-19T19:40:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.27205v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T04:48:33Z", "role": "primary", "sha256": "beefe28f2bbc7211f8c2a7f01fe23d88b97e9769059f48d885068566a88188ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22+OR+abs%3A%22web+navigation%22+OR+abs%3AAndroidWorld+OR+abs%3AMind2Web%29+AND+submittedDate%3A%5B202610070000+TO+202610210000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.27205", "identifiers": {"arxiv": "2610.27205"}, "kind": "paper", "last_verified_at": "2026-10-21T04:48:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.27205"}], "name": "LoWAgent: A Low-Resource Multilingual Benchmark for Web Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.27205"}
{"aliases": [], "authors": ["Xu Hu", "Haoming Li", "Yuxiang Wang", "Jinxin Yang", "Kangshuo Li", "Licheng Wang", "Dingyi Kang", "Jiayu Song", "Tianhao Wang", "Zhanhua Hu", "Xu Yang", "Qiannan Li", "Bingzhe Li", "Feng Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8a04fe46937c3a771a10373108d6979469a1b56aac823a3afc64e01deb8241a", "submitted_at": "2026-10-20T05:06:00Z", "updated_at": "2026-10-20T05:06:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.27645v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T04:48:33Z", "role": "primary", "sha256": "beefe28f2bbc7211f8c2a7f01fe23d88b97e9769059f48d885068566a88188ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22+OR+abs%3A%22web+navigation%22+OR+abs%3AAndroidWorld+OR+abs%3AMind2Web%29+AND+submittedDate%3A%5B202610070000+TO+202610210000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.27645", "identifiers": {"arxiv": "2610.27645"}, "kind": "paper", "last_verified_at": "2026-10-21T04:48:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.27645"}], "name": "CUA-Recovery: Benchmarking Error Recovery in Long-Horizon Computer-Use Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-20", "url": "https://arxiv.org/abs/2610.27645"}
{"aliases": [], "authors": ["Haocheng Wang", "Youquan Liu", "Weiyu Guo", "Aokai Zhang", "Donghang Wu", "Hongyu Liu", "Sha Zhang", "Weidong Yang", "Ben Fei"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecd5f01ccdc4fa4875768b72c15d77b68b6cabd99ef6d9125f3462ae1af12949", "submitted_at": "2026-10-20T08:42:42Z", "updated_at": "2026-10-20T08:42:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.27861v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-22T13:49:08Z", "role": "primary", "sha256": "520674c9e9007807e4df0257d492cfbbf8a36b8a547b1a812af7ac3265fb3e02", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.27861&max_results=1"}], "id": "arxiv:2610.27861", "identifiers": {"arxiv": "2610.27861"}, "kind": "paper", "last_verified_at": "2026-10-22T13:49:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.27861"}], "name": "RoboCasa-Shift: Stage-Aware Robustness Evaluation of Vision-Language-Action Policies in Long-Horizon Household Manipulation", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-20", "url": "https://arxiv.org/abs/2610.27861"}
{"aliases": [], "authors": ["Byeonghun Bang", "Sangwoo Park", "Jinheon Baek", "Kangsan Kim", "Youngwan Lee", "Sung Ju Hwang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "566cdf36884edceec5842557c775396da69c25049ffd6f7498c9043b16dfac79", "submitted_at": "2026-10-20T09:59:12Z", "updated_at": "2026-10-20T09:59:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.27954v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T04:48:33Z", "role": "primary", "sha256": "beefe28f2bbc7211f8c2a7f01fe23d88b97e9769059f48d885068566a88188ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22+OR+abs%3A%22web+navigation%22+OR+abs%3AAndroidWorld+OR+abs%3AMind2Web%29+AND+submittedDate%3A%5B202610070000+TO+202610210000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.27954", "identifiers": {"arxiv": "2610.27954"}, "kind": "paper", "last_verified_at": "2026-10-21T04:48:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.27954"}], "name": "Lost in Recovery: Demystifying Recovery Behavior of Computer-Use Agents via Goal-Oriented Action Replay", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-20", "url": "https://arxiv.org/abs/2610.27954"}
{"aliases": [], "authors": ["Saaket Agashe", "Yuzhe Yang", "Gonzalo Gonzalez-Pumariega", "Bhuvan Chandra Koduru", "Chih-Lun Li", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b8b3a8a83d7fca0f8950110b74ffe51c11bb7d403677d4a822bc7597dea211e9", "submitted_at": "2026-10-20T17:47:37Z", "updated_at": "2026-10-20T17:47:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.28601v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T04:48:33Z", "role": "primary", "sha256": "beefe28f2bbc7211f8c2a7f01fe23d88b97e9769059f48d885068566a88188ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22+OR+abs%3A%22web+navigation%22+OR+abs%3AAndroidWorld+OR+abs%3AMind2Web%29+AND+submittedDate%3A%5B202610070000+TO+202610210000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.28601", "identifiers": {"arxiv": "2610.28601"}, "kind": "paper", "last_verified_at": "2026-10-21T04:48:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.28601"}], "name": "Evolving Agent Interfaces", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-20", "url": "https://arxiv.org/abs/2610.28601"}
{"aliases": [], "authors": ["Jingyu Xiao", "Dasen Dai", "Yuxuan Wan", "Jen-tse Huang", "Yintong Huo", "Hongyu Zhang", "Michael R. Lyu"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d674ff8ea7ca3ce481e0c1d595a1f13fd0079d6a4f2752ed8bc6aedfea3db220", "submitted_at": "2026-10-20T17:53:24Z", "updated_at": "2026-10-20T17:53:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.28610v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-21T04:48:33Z", "role": "primary", "sha256": "beefe28f2bbc7211f8c2a7f01fe23d88b97e9769059f48d885068566a88188ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22+OR+abs%3A%22web+navigation%22+OR+abs%3AAndroidWorld+OR+abs%3AMind2Web%29+AND+submittedDate%3A%5B202610070000+TO+202610210000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.28610", "identifiers": {"arxiv": "2610.28610"}, "kind": "paper", "last_verified_at": "2026-10-21T04:48:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.28610"}], "name": "InterWebCoder: Learning Interactive Webpage Generation from Executable Tests via Reinforcement Learning", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-20", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-20", "url": "https://arxiv.org/abs/2610.28610"}
{"aliases": [], "authors": ["Jianing Yin", "Ziyi Chen", "Hao Wang", "Yiyou Sun", "Zijian Jin", "Dawn Song", "Daoan Zhang"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2036dfb06df824248ec9fd2e6088bbef574dbc5c380a0f33a20b2fc382c988b5", "submitted_at": "2026-10-19T17:22:36Z", "updated_at": "2026-10-19T17:22:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.28679v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-22T20:07:57Z", "role": "primary", "sha256": "d4ba11fca86dad31e5c56743a336d0d56442ff79791d124327964d7e13a21548", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.28679&max_results=1"}], "id": "arxiv:2610.28679", "identifiers": {"arxiv": "2610.28679"}, "kind": "paper", "last_verified_at": "2026-10-22T20:07:57Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.28679"}], "name": "VerifierBench: Benchmarking the Generation of Executable Verifiers for LLM Agents", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-19", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-19", "url": "https://arxiv.org/abs/2610.28679"}
{"aliases": [], "authors": ["Yunfeng Zhu", "Shengcheng Yu", "Zihan Li", "Zhenyu Chen"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "549ebe784b04c1acde6875e369708b1b566d449b81b6fea38d2e9cb6ee11a11c", "submitted_at": "2026-10-21T11:50:17Z", "updated_at": "2026-10-21T11:50:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.29609v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-22T04:46:44Z", "role": "primary", "sha256": "68c5c99210c4150b1c361241e24caf254eddfd302e1b2152db9907417c46e650", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22+OR+abs%3A%22web+navigation%22+OR+abs%3AAndroidWorld+OR+abs%3AMind2Web%29+AND+submittedDate%3A%5B202610080000+TO+202610220000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.29609", "identifiers": {"arxiv": "2610.29609"}, "kind": "paper", "last_verified_at": "2026-10-22T04:46:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.29609"}], "name": "Who Verifies the Agent's Success Claim? Self-Report Leakage in GUI-Agent Completion Judging", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-21", "url": "https://arxiv.org/abs/2610.29609"}
{"aliases": [], "authors": ["Yanzhe Chen", "Ziyi Yang", "Jifeng Zhu", "Qiming Huang", "Ruihe An", "Peiyao Xu", "Hesen Yang", "Runda Liu", "Chang Gong", "Zhijun Cao", "Zechen Bai", "Wenzheng Zeng", "Yiqi Lin", "Guoqiang Liang", "Kevin Yuchen Ma", "Kevin Qinghong Lin", "Mike Zheng Shou"], "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5be9e96cd0e1b7574aa96426b1bff6e6f93db605a2e803e39b0f2314d2293fb7", "submitted_at": "2026-10-21T17:52:48Z", "updated_at": "2026-10-21T17:52:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.30108v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-22T04:46:44Z", "role": "primary", "sha256": "68c5c99210c4150b1c361241e24caf254eddfd302e1b2152db9907417c46e650", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22+OR+abs%3A%22web+navigation%22+OR+abs%3AAndroidWorld+OR+abs%3AMind2Web%29+AND+submittedDate%3A%5B202610080000+TO+202610220000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}], "id": "arxiv:2610.30108", "identifiers": {"arxiv": "2610.30108"}, "kind": "paper", "last_verified_at": "2026-10-22T04:46:44Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "paper", "url": "https://arxiv.org/abs/2610.30108"}], "name": "Multimodal Embodied Agents: A Unified Capability-centric Perspective from Computer-Use to Robot-Use", "notes": "Bibliographic metadata checked against arXiv. Relevance and taxonomy await editorial review; presence is not an endorsement.", "published_at": "2026-10-21", "relations": [], "review_status": "metadata_verified", "scope": "relevance_pending", "summary": null, "topics": [], "updated_at": "2026-10-21", "url": "https://arxiv.org/abs/2610.30108"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md", "population": "AndroidControl collection", "qualifier": "more-than", "unit": "demonstrations", "value": 15000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md", "population": "AndroidControl collection", "qualifier": "exact", "unit": "applications", "value": 833}], "evidence_urls": ["https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md"], "formats": ["GZIP-compressed TFRecord"], "modalities": ["screenshots", "accessibility trees", "instructions", "JSON actions"], "notes": "Step instructions can be repeated across multiple actions. Each episode has one more screenshot than action; do not count repeated step instructions as independent task goals.", "splits": ["publisher split file"], "supervision": ["interaction trajectories", "high- and low-level action instruction"]}, "modalities": ["screenshot", "accessibility-tree", "instruction", "action"], "reported_scale": {"comparison": ">", "unit": "demonstrations", "value": 15000}, "sample_count": null, "sample_unit": "demonstrations"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:29Z", "role": "primary", "sha256": "e4f8baa15441fb63cad07220cc5a4de177a1f785d4fe1fed5cfcbf3c3867faf1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.reported_scale", "details.modalities", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/google-research/google-research/HEAD/LICENSE"}], "id": "component:google-research/android_control", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:29Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/google-research/google-research/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research/google-research/tree/master/android_control"}], "name": "AndroidControl", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md", "target_id": "arxiv:2406.03679", "type": "introduced_by"}], "review_status": "curated", "scope": "core", "summary": "Android demonstrations with high-level goals, step instructions, screenshots, and accessibility trees.", "topics": ["data-generation", "mobile-navigation"], "updated_at": null, "url": "https://github.com/google-research/google-research/tree/master/android_control"}
{"aliases": [], "authors": [], "details": {"access": "public download referenced by publisher", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md", "population": "sum of the five published AitW subsets", "qualifier": "exact", "unit": "episodes", "value": 715142}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md", "population": "sum of the five published AitW subsets", "qualifier": "exact", "unit": "action examples", "value": 5689993}], "evidence_urls": ["https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md"], "formats": ["GZIP-compressed TFRecord"], "modalities": ["screenshots", "instructions", "actions"], "notes": "Episodes and action examples are different units. The collection combines multi-step subsets and hindsight-labeled single-step tasks; preserve subset and split choices.", "splits": [], "supervision": ["interaction trajectories", "action prediction"]}, "download_url": "https://console.cloud.google.com/storage/browser/gresearch/android-in-the-wild", "example_count": 5689993, "modalities": ["screenshot", "action", "instruction"], "sample_count": 715142, "sample_unit": "episodes"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:29Z", "role": "primary", "sha256": "6d8964c95db41dada1d5867e50a5ec61f6cb90595965d0e72e010b8e4a12dcb2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.example_count", "details.modalities", "details.access", "details.download_url", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/google-research/google-research/HEAD/LICENSE"}], "id": "component:google-research/android_in_the_wild", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:29Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/google-research/google-research/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research/google-research/tree/master/android_in_the_wild"}], "name": "Android in the Wild", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md", "target_id": "arxiv:2307.10088", "type": "introduced_by"}], "review_status": "curated", "scope": "core", "summary": "Human demonstrations of Android tasks with screenshots and touch actions.", "topics": ["data-generation", "mobile-navigation"], "updated_at": null, "url": "https://github.com/google-research/google-research/tree/master/android_in_the_wild"}
{"aliases": [], "authors": [], "details": {"component_path": "WebEvolver", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/License.txt"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "a82450e1b5e15070bee79ee8b79f8d9a97cb85f7", "repository_url": "https://github.com/Tencent/SelfEvolvingAgent"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "447dfa1e52789f7f3f111e9e0d453fb1f9fd8bf2d7b858f28fbe21a2239b4b2e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/WebEvolver/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:01Z", "role": "primary", "sha256": "4214af42bcadf1ea52a46c2008f8e741f0aa955a58f346d9880858589be7de1d", "supports": ["identity"], "url": "https://api.github.com/repos/Tencent/SelfEvolvingAgent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:25Z", "role": "primary", "sha256": "ad0f4ae4b58b5a0e84feeb5df717929b97ea37cb945c17c98510ba54509c99e7", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/Tencent/SelfEvolvingAgent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:13Z", "role": "primary", "sha256": "873a78a46309fa3892d94b72cf73fd55542bae0e0ba8235afb26a848a64a8255", "supports": ["identity"], "url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "1d7a5369af7ec336bc173dba6cfee9ad0898e527f96c5b7a10cee930806c67a3", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/License.txt"}], "id": "component:tencent/selfevolvingagent/webevolver", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:14Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/License.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Tencent/SelfEvolvingAgent/tree/main/WebEvolver"}], "name": "WebEvolver code component", "notes": "Scoped to the WebEvolver directory; the repository also hosts WebCoT and VScan, which are not merged into this resource. Training and model services require separate setup. Root License.txt explicitly declares MIT for SelfEvolvingAgent and separately identifies CogKernel’s Apache-2.0 dependency terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/WebEvolver/README.md", "target_id": "arxiv:2504.21024", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Component for web-agent trajectory collection, supervised-data preparation and world-model-based trajectory synthesis.", "topics": ["web-navigation", "world-models", "data-generation", "self-reflection"], "updated_at": null, "url": "https://github.com/Tencent/SelfEvolvingAgent/tree/main/WebEvolver"}
{"aliases": [], "authors": [], "details": {"component_path": "Mobile-Agent-E", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "11cea575561fb7800b5fb6b6cafa56f7a91de11f", "repository_url": "https://github.com/X-PLUG/MobileAgent"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:15Z", "role": "primary", "sha256": "f8515fbce73ea0ef3d1d558d4567bb949244e6384e61c956ba3ddbc27294f0ae", "supports": ["name", "kind", "environments", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:03Z", "role": "primary", "sha256": "d403251020124bebcd9f180f800c8ea9f18a3f117c82fcf04c3cbe9ffe0e35f9", "supports": ["identity"], "url": "https://api.github.com/repos/X-PLUG/MobileAgent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:21Z", "role": "primary", "sha256": "7293efadde7c96b64ff8483b2157241e17d0b0fbfff4839de9c7b11485b07cbb", "supports": ["identity", "release_status", "details.repository_tree_sha", "relations"], "url": "https://api.github.com/repos/X-PLUG/MobileAgent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["summary", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "961819a89dacd1f4f161e1d2ebf616b538bf095a2eecfa3ef40b7ccec633fac4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}], "id": "component:x-plug/mobileagent/mobile-agent-e", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-18T04:27:15Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/X-PLUG/MobileAgent/tree/main/Mobile-Agent-E"}], "name": "Mobile-Agent-E code component", "notes": "The inspected README explicitly supports Android only, with experiments on Samsung Galaxy A15; host-machine macOS/Linux instructions do not imply control of those operating systems. Agent code, launch scripts and benchmark definitions are present. The root MIT code license does not establish terms for the separately published task dataset or external models.", "published_at": null, "relations": [{"evidence_url": "https://api.github.com/repos/X-PLUG/MobileAgent/git/trees/main?recursive=1", "target_id": "github:x-plug/mobileagent", "type": "component_of"}, {"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md", "target_id": "arxiv:2501.11733", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Android automation component with hierarchical planning, reflection and persistent tips/shortcuts learned across task sequences.", "topics": ["planning", "memory", "self-reflection", "mobile-navigation"], "updated_at": null, "url": "https://github.com/X-PLUG/MobileAgent/tree/main/Mobile-Agent-E"}
{"aliases": [], "authors": [], "details": {"component_path": "PC-Agent", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "11cea575561fb7800b5fb6b6cafa56f7a91de11f", "repository_url": "https://github.com/X-PLUG/MobileAgent"}, "environments": ["desktop", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:15Z", "role": "primary", "sha256": "ca5b488f10c229b79cc6f9e1f32007d7d84c34366492a10b90d0d66622e77273", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/PC-Agent/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:03Z", "role": "primary", "sha256": "d403251020124bebcd9f180f800c8ea9f18a3f117c82fcf04c3cbe9ffe0e35f9", "supports": ["identity"], "url": "https://api.github.com/repos/X-PLUG/MobileAgent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:21Z", "role": "primary", "sha256": "7293efadde7c96b64ff8483b2157241e17d0b0fbfff4839de9c7b11485b07cbb", "supports": ["identity", "release_status", "details.repository_tree_sha", "relations"], "url": "https://api.github.com/repos/X-PLUG/MobileAgent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "961819a89dacd1f4f161e1d2ebf616b538bf095a2eecfa3ef40b7ccec633fac4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}], "id": "component:x-plug/mobileagent/pc-agent", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-18T04:27:15Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/X-PLUG/MobileAgent/tree/main/PC-Agent"}], "name": "PC-Agent (X-PLUG component)", "notes": "Scoped to PC-Agent in the MobileAgent monorepo and distinct from GAIR-NLP/PC-Agent. The source tree contains current and v1 implementations; the README states a March 2025 code update. Root MIT covers repository code; the separately linked PC-Eval tasks and application software are not assigned that license by this record.", "published_at": null, "relations": [{"evidence_url": "https://api.github.com/repos/X-PLUG/MobileAgent/git/trees/main?recursive=1", "target_id": "github:x-plug/mobileagent", "type": "component_of"}, {"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/PC-Agent/README.md", "target_id": "arxiv:2502.14282", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Multi-agent desktop automation component for Windows and macOS, combining active perception, task decomposition and reflection for workflows spanning applications.", "topics": ["planning", "cross-app", "self-reflection"], "updated_at": null, "url": "https://github.com/X-PLUG/MobileAgent/tree/main/PC-Agent"}
{"aliases": [], "authors": [], "details": {"availability": "Public Web Agent source and Chrome-extension archive manifest with local setup instructions.", "component_path": "real_agents/web_agent", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/LICENSE"], "reason": "OpenAgents source license, scoped here to the reviewed Web Agent component; bundled third-party components retain their notices."}}, "repository_revision": "ff2e46440699af1324eb25655b622c4a131265bb"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "fb6ef7fdb133612467e27ffa642ccbe73cfb4bd8fb0c63c1bc629ad686694e1d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "b843b0519dc74bf6dbf693e0274e8ef8eb1ade94a656d3ab23faac2df4f7c9fa", "supports": ["identity"], "url": "https://api.github.com/repos/xlang-ai/OpenAgents"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "cf5a65d5a6b7ba2a9cf950421f6902534d523f7e8d850d1b23bd5395fd3d04c8", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/xlang-ai/OpenAgents/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "bcd2232e1ec4ad9ba0628cddd30031c364d62c6f112d8ecc2c24b9a4d35696af", "supports": ["release_status", "notes", "environments"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/frontend/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:45:23Z", "role": "primary", "sha256": "5b7084cfe6149602740cf4cd85e958142aed3aea0a0a0345a82ec1618ceb7b19", "supports": ["summary", "environments", "scope"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/real_agents/web_agent/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:09Z", "role": "primary", "sha256": "a86eea1027b8d92a68ded4231671379388df757992fa69d3e8fc8f8115d124fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/LICENSE"}], "id": "component:xlang-ai/openagents/web_agent", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": {"basis": "OpenAgents source license, scoped here to the reviewed Web Agent component; bundled third-party components retain their notices.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T12:34:09Z", "url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/OpenAgents/tree/main/real_agents/web_agent"}], "name": "OpenAgents Web Agent", "notes": "This component record covers real_agents/web_agent and the documented WeBot Chrome-extension integration. The broader Data Agent and Plugins Agent are not treated as GUI-control assets. The source manifest includes the agent implementation and frontend/webot_extension.zip; the archive was not downloaded or executed. The README describes Docker deployment as under development and recommends source deployment. Apache-2.0 applies to project source; third-party extension/dependency notices remain separate.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/README.md", "target_id": "arxiv:2310.10634", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The OpenAgents browser-control component, connecting a Chrome extension to a web agent and local chat platform for navigating sites and completing interface tasks.", "topics": ["web-navigation", "automation", "tool-integration"], "updated_at": null, "url": "https://github.com/xlang-ai/OpenAgents/tree/main/real_agents/web_agent"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:05Z", "role": "primary", "sha256": "86d7fbe30596de4a1f1a1df2a986f5f247351a91c0c593dc4c29ba9c70d28c8f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/aaronlab/browsertrace/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "7b4ea96686420061c9d0e1ad4984d7c5efb7e1a62ce6dffd47862f02a7260867", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/aaronlab/browsertrace/HEAD/LICENSE"}], "id": "github:aaronlab/browsertrace", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:05Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/aaronlab/browsertrace/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/aaronlab/browsertrace"}], "name": "BrowserTrace", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Local, cloud-free replay debugger that records each Browser Use step's screenshot, URL, action and model I/O as a timeline and reports the step at which a failed run diverges from a known-good one.", "topics": ["observability"], "updated_at": null, "url": "https://github.com/aaronlab/browsertrace"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/aburns4/MoTIF/HEAD/README.md"], "formats": ["JSON metadata", "screen images"], "modalities": ["screenshots", "view hierarchies", "language", "interaction sequences", "feasibility labels"], "notes": "The raw collection mixes action, language, exploration, and unusable/other traces. These groups have different annotation coverage; not every raw trace is a valid task demonstration.", "splits": [], "supervision": ["interaction trajectories", "task feasibility"]}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "4b2f5a23dc28e96b671272380b953e7fabaf34f4b1e83f3513643846718e3389", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/aburns4/MoTIF/HEAD/README.md"}], "id": "github:aburns4/motif", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/aburns4/MoTIF"}], "name": "MoTIF", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/aburns4/MoTIF/HEAD/README.md", "target_id": "arxiv:2202.02312", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile task demonstrations and natural-language instructions.", "topics": ["mobile-navigation"], "updated_at": null, "url": "https://github.com/aburns4/MoTIF"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["mobile GUI operations"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending", "population": "AppSim-Bench paper-described release", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2609.07712", "unit": "tasks", "value": 557}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Acrab-Agentic-Labs/AppSim/HEAD/README.md", "population": "published simulated app set", "qualifier": "exact", "unit": "applications", "value": 17}], "evaluation": "Controlled backend data and outcome-based task verification.", "evidence_urls": ["https://raw.githubusercontent.com/Acrab-Agentic-Labs/AppSim/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"], "focus": ["mobile task completion"], "notes": "Use the simulated apps and controlled backend state supplied by the benchmark. Desktop host instructions are not the platform being evaluated.", "observations": ["mobile interfaces"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "12dc2afe339c1ed66e4a6a4ccf5ae13d27ce32f452960341529a4b9016f69059", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/Acrab-Agentic-Labs/AppSim/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "github:acrab-agentic-labs/appsim", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Acrab-Agentic-Labs/AppSim"}], "name": "AppSim-Bench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Acrab-Agentic-Labs/AppSim/HEAD/README.md", "target_id": "arxiv:2609.07712", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Simulated mobile applications and evaluation tooling for reproducible GUI-agent tasks with controlled application state.", "topics": ["evaluation", "sandbox"], "updated_at": null, "url": "https://github.com/Acrab-Agentic-Labs/AppSim"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "0fe34cd605e492c4c93f760e9caefd194f9c8ec7cfc3d4d4a927a1ac1613e51c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/agent-infra/sandbox/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/agent-infra/sandbox/HEAD/LICENSE"}], "id": "github:agent-infra/sandbox", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/agent-infra/sandbox/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/agent-infra/sandbox"}], "name": "Agent Infra Sandbox", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Sandbox environments supporting computer and browser interaction.", "topics": ["sandbox"], "updated_at": null, "url": "https://github.com/agent-infra/sandbox"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "3ee6212ef10c39943546d20a3beeed412262f4a8a910120becb5409626569b0f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/agentsea/agentdesk/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "40f2354eb662ebb7b99d874b45f639e1265bec931073cda9ff21c6cf2cb0c499", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/agentsea/agentdesk/HEAD/LICENSE"}], "id": "github:agentsea/agentdesk", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/agentsea/agentdesk/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/agentsea/agentdesk"}], "name": "AgentDesk", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Desktop environments that AI agents control programmatically, run locally with Docker or remotely on Kubernetes, built on the agentd REST daemon and offering a coordinate-grid overlay for LMM input.", "topics": ["sandbox"], "updated_at": null, "url": "https://github.com/agentsea/agentdesk"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "32420db547a889f0d86090689d0bd0c93d1776a70346dfd4029db7d99ae4bd9f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/agentsea/surfkit/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "40f2354eb662ebb7b99d874b45f639e1265bec931073cda9ff21c6cf2cb0c499", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/agentsea/surfkit/HEAD/LICENSE"}], "id": "github:agentsea/surfkit", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/agentsea/surfkit/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/agentsea/surfkit"}], "name": "Surfkit", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Python toolkit and CLI for building, publishing and running multimodal agents that operate devices such as Ubuntu desktops, locally or on Kubernetes, with task trackers and an agent hub.", "topics": ["workflow"], "updated_at": null, "url": "https://github.com/agentsea/surfkit"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["agent-specific Android actions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "population": "complete README task collection", "qualifier": "exact", "unit": "tasks", "value": 340}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "population": "single-app subset", "qualifier": "exact", "unit": "tasks", "value": 300}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "population": "cross-app subset", "qualifier": "exact", "unit": "tasks", "value": 40}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "population": "README total including third-party, Google, and system apps", "qualifier": "exact", "unit": "applications", "value": 66}], "evaluation": "A multimodal evaluator and annotated key-component matching assess completion; the framework also reports resource-use metrics.", "evidence_urls": ["https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/Documentation.md"], "focus": ["mobile task completion", "resource use", "cross-app workflows"], "notes": "The two task subsets partition the 340 tasks. The integrated agent and evaluator modes must accompany results. The preconfigured emulator snapshot is still marked coming soon in the inspected documentation.", "observations": ["screenshot trajectories"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "5294a67728072a3658cb2073bdf9a084fe3c00d87a848923dff68adb35213072", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "5294a67728072a3658cb2073bdf9a084fe3c00d87a848923dff68adb35213072", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:23Z", "role": "primary", "sha256": "a55290d62f8823097c73d74bac4eb08853a3caff724382e60a6a9364e7ced987", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/Documentation.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:58Z", "role": "primary", "sha256": "6962bfc8ee16c6e985e4fb680b8fe46d5b34d30f7cb8a1e61fd4c55c93120ddf", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/LICENSE"}], "id": "github:ai-agents-2030/spa-bench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:58Z", "url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ai-agents-2030/SPA-Bench"}], "name": "SPA-Bench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "target_id": "arxiv:2410.15164", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "340 single-app and cross-app tasks over 66 Chinese and English Android apps, with plug-in support for 11 smartphone agents and automatic coarse-to-fine success detection.", "topics": ["evaluation", "mobile-navigation"], "updated_at": null, "url": "https://github.com/ai-agents-2030/SPA-Bench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:07Z", "role": "primary", "sha256": "b68b2cc26bfc6a654fec35f3c53c1b1334f3d617671673407d3ed3b59f3db973", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ai-nikolai/StateAct/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "38f8c08395c6e67ff9bf52b27e772667f193a172bd6f726507c57776ad3e5e5f", "supports": ["identity"], "url": "https://api.github.com/repos/ai-nikolai/StateAct"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:07Z", "role": "primary", "sha256": "97841e6eab1002bcea949e031f5abd3567482ebbcf03b991bd64643027c468bf", "supports": ["release_status"], "url": "https://api.github.com/repos/ai-nikolai/StateAct/git/trees/main?recursive=1"}], "id": "github:ai-nikolai/stateact", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:02:07Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ai-nikolai/StateAct"}], "name": "StateAct", "notes": "The repository contains a WebShop runner as well as ALFWorld and TextCraft experiments. Inclusion concerns its web-shopping interaction component. No code license or downloadable trained checkpoint is declared in the inspected root metadata.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ai-nikolai/StateAct/HEAD/README.md", "target_id": "arxiv:2410.02810", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "State-tracking prompts and experiment runners for language agents, including a WebShop implementation alongside text-based environments.", "topics": ["planning", "state-tracking", "web-navigation"], "updated_at": null, "url": "https://github.com/ai-nikolai/StateAct"}
{"aliases": ["AdvWeb"], "authors": [], "details": {"availability": "Public research source and workflow documentation; no standalone trained checkpoint was verified.", "repository_revision": "e01316e33d11ab145daa7cc7ec12d6ddb0bf38a0"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:18:18Z", "role": "primary", "sha256": "019d45dd4c383f919f3ac2bf84c672500d63751a49de2dd2bac9c018ffc99881", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/AI-secure/AdvAgent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:18:17Z", "role": "primary", "sha256": "3ea44a202a677c0beb744ab05bd687da93a83365b7eb0b7999c3a4c22d7e2608", "supports": ["identity"], "url": "https://api.github.com/repos/AI-secure/AdvAgent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:18:33Z", "role": "primary", "sha256": "95929797eb0717ce26b26ccbdc752bfda18fc0070975b32f09827b4e9fec6ecf", "supports": ["release_status", "licenses.code"], "url": "https://api.github.com/repos/AI-secure/AdvAgent/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:17:48Z", "role": "primary", "sha256": "3ea44a202a677c0beb744ab05bd687da93a83365b7eb0b7999c3a4c22d7e2608", "supports": ["identity"], "url": "https://api.github.com/repos/AI-secure/AdvWeb"}], "id": "github:ai-secure/advagent", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T09:18:18Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AI-secure/AdvAgent"}], "name": "AdvAgent (formerly AdvWeb)", "notes": "The official GitHub API resolves AI-secure/AdvWeb to AI-secure/AdvAgent. Public data-generation, SFT/DPO training and evaluation scripts are present, with upstream data and external model APIs required. No repository license was found, and no permission for third-party data or model weights is inferred. The README citation block is empty, so no paper identifier is bound from title similarity alone.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Research code for constructing and evaluating adversarial inputs to web agents, with a black-box red-team pipeline using Mind2Web data and SeeAct.", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": null, "url": "https://github.com/AI-secure/AdvAgent"}
{"aliases": [], "authors": [], "details": {"availability": "Public Android automation source and setup instructions; external models/services required.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/LICENSE"], "reason": "Repository GNU GPL version 3 license; pretrained models and linked data retain separate terms."}}, "repository_revision": "d31b50e98bde28c0d1cfbc7d7ee2528d10b3fbb0"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:18Z", "role": "primary", "sha256": "816981b343628c53188fd27041d62f52854fcfe83c976a125b84545f5e010274", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:23Z", "role": "primary", "sha256": "94d8e95d302160e3c581015267dcfbb9f61c41ad1f04e470fcd71f5ee1ffa25d", "supports": ["identity"], "url": "https://api.github.com/repos/AkimotoAyako/VisionTasker"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:48Z", "role": "primary", "sha256": "54a1c08d44e0ad1d60bee57e3db2842518e014db3be8b08362e484315f3a6a6f", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/AkimotoAyako/VisionTasker/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:03Z", "role": "primary", "sha256": "3972dc9744f6499f0f9b2dbf76696f2ae7ad8af9b23dde66d6af86c9dfb36986", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/LICENSE"}], "id": "github:akimotoayako/visiontasker", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T11:20:18Z", "licenses": {"code": {"basis": "Repository GNU GPL version 3 license; pretrained models and linked data retain separate terms.", "declared": "GPL-3.0", "retrieved_at": "2026-10-17T11:23:03Z", "url": "https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AkimotoAyako/VisionTasker"}], "name": "VisionTasker", "notes": "The official lowercase readme.md and source tree expose automatic and step-by-step entry points. The documented setup uses a Windows host and an Android phone connected through ADB; it also requires separately obtained detection/CLIP models and model/OCR services. GPL-3.0 is recorded for the code only. External model download links were not used to infer weight licenses or verify a separate checkpoint release.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/readme.md", "target_id": "arxiv:2312.11190", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An Android automation implementation that translates screenshot-based UI understanding into language descriptions for step-by-step LLM planning and action execution.", "topics": ["mobile-navigation", "screen-understanding", "planning"], "updated_at": null, "url": "https://github.com/AkimotoAyako/VisionTasker"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "e06aedcb1760867470afeb27d4eb4cdde9a3b76b9d55af183047df3c0ae2b71c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/alibaba/UI-Ins/HEAD/README.md"}], "id": "github:alibaba/ui-ins", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/alibaba/UI-Ins"}], "name": "UI-Ins", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/alibaba/UI-Ins/HEAD/README.md", "target_id": "arxiv:2510.20286", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Treats grounding instructions as multiple reasoning perspectives, releasing a data pipeline, training code and UI-Ins-7B and 32B models trained by SFT on synthesized instructions, then RL.", "topics": ["data-generation"], "updated_at": null, "url": "https://github.com/alibaba/UI-Ins"}
{"aliases": [], "authors": [], "details": {"availability": "announced; data and sampling pipelines in preparation", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md", "population": "paper-described synthesis output; not a released file inventory", "qualifier": "approximately", "unit": "training samples", "value": 800000}], "evidence_urls": ["https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md"], "formats": [], "modalities": ["GUI interaction trajectories"], "notes": "The inspected README marks the dataset and collection pipelines as in preparation. This profile records an announced research asset, not a verified downloadable corpus; file formats, splits, and release-specific data terms remain unknown.", "splits": [], "supervision": ["recovery trajectory synthesis"]}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "bf5c1469d33ced951ff0a5de017494b65f654023ceee8e7fbfd8b60afb082244", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "bf5c1469d33ced951ff0a5de017494b65f654023ceee8e7fbfd8b60afb082244", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "bf5c1469d33ced951ff0a5de017494b65f654023ceee8e7fbfd8b60afb082244", "supports": ["summary", "notes", "details.availability"], "url": "https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md"}], "id": "github:alibabaresearch/rots", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AlibabaResearch/RoTS"}], "name": "RoTS", "notes": "The 800k figure describes training samples in the study, not an independently counted public download. The repository code-license declaration does not establish terms for an unreleased dataset.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Announced GUI trajectory synthesis resources for recovery from policy-induced errors; the inspected repository marks code and data releases as in preparation.", "topics": ["data-generation", "robustness"], "updated_at": null, "url": "https://github.com/AlibabaResearch/RoTS"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "c4da2b089f97296ac6c7b3ef7273db57b650d12cf3595f6f6ea74c901e5277b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/alipay/mobile-agent/HEAD/README.md"}], "id": "github:alipay/mobile-agent", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/alipay/mobile-agent"}], "name": "MobileAgent (Alipay)", "notes": "Distinct from the X-PLUG/MobileAgent project despite the similar name.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/alipay/mobile-agent/HEAD/README.md", "target_id": "arxiv:2401.04124", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile-control research implementation combining operating-procedure context with human interaction for personalized task execution.", "topics": ["mobile-navigation", "workflow", "human-in-the-loop"], "updated_at": null, "url": "https://github.com/alipay/mobile-agent"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:14:29Z", "role": "primary", "sha256": "281793fe6b875bd1a63591502345ac9523b1127acd9f41d5c2c88933a357e4f2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Alokia/COLA-demo/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["relations", "environments"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "github:alokia/cola-demo", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T13:14:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Alokia/COLA-demo"}], "name": "COLA demo", "notes": "The publisher describes this as demonstration code and states that the full implementation is not released.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50", "target_id": "arxiv:2503.09263", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Research demonstration of collaborative Windows GUI automation with dynamic task scheduling and recovery mechanisms.", "topics": ["automation", "planning", "human-in-the-loop"], "updated_at": null, "url": "https://github.com/Alokia/COLA-demo"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "3e5f24b3175afeba8635dacc7b66eba1ce89fea49996a20281ff2df5a41f1e24", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/amazon-science/PAE/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "ae4f966004dd672fd5b994add5a5cba1fee9233d4d81b224e94c339728fafd87", "supports": ["identity"], "url": "https://api.github.com/repos/amazon-science/PAE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "b1e24a3f38fae67f71033a3f920164d282d3f8bb072f438349dbe9eb23730a7a", "supports": ["release_status"], "url": "https://api.github.com/repos/amazon-science/PAE/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:15Z", "role": "primary", "sha256": "db2e35513dbadcdc67f5819a3bfee2777786538dd3531620cd5fbd4b6ed6e538", "supports": ["licenses"], "url": "https://raw.githubusercontent.com/amazon-science/PAE/HEAD/LICENSE.txt"}], "id": "github:amazon-science/pae", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/amazon-science/PAE"}], "name": "Proposer-Agent-Evaluator (PAE)", "notes": "Source code and browser-environment modules are present. The repository LICENSE is CC-BY-NC-4.0, while its README describes that license as applying to a dataset, leaving code scope ambiguous. The separately linked dataset is gated and was not accessed; model and dataset terms must be checked separately.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/amazon-science/PAE/HEAD/README.md", "target_id": "arxiv:2412.13194", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A browser-agent training system that proposes tasks, collects visual interaction trajectories, and uses automated evaluation to drive policy improvement.", "topics": ["reinforcement-learning", "web-navigation", "data-generation"], "updated_at": null, "url": "https://github.com/amazon-science/PAE"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["app launch/stop", "click/double-click/long-click", "text input", "swipe/drag and Android navigation", "finish"], "counts": [], "evaluation": "The released evaluator compares trajectories with reference actions using longest-common-subsequence-based task reward and completion. Its reversed redundancy metric is the reference/predicted action-length ratio (with 1e-6 denominator smoothing), multiplied by the LLM success flag. LLM success judgments and instruction-constraint checks are separate metrics. Full reference-dependent evaluation coverage cannot be assumed because complete annotated action sequences are not released.", "evidence_urls": ["https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/README.md", "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/README.md", "https://api.github.com/repos/AndroidArenaAgent/AndroidArena/git/trees/main?recursive=1", "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/observation.py", "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/actions.py", "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/run_evaluator.py"], "focus": ["multi-step Android tasks", "cross-app workflows and instruction constraints"], "notes": "README states that task instructions are released but annotated action sequences, apart from the calendar example, will come later. Environment setup targets Android API 33 with uiautomator2 and manual Google-account setup. Public task YAML and evaluator code exist, but benchmark execution and full reference coverage were not verified. No task total is derived from app YAML file counts. Code and data licenses remain unknown.", "observations": ["UI hierarchy text", "screenshots", "combined text/image observations"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "206bd0dc5220cf754cffc97767cce6a3cd27660ef050a569366bef2331241c60", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "a1e7dc0c5facae030c2cf19aad6511996e7e1198d8d6619268d10c29a1b57bcf", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "edc03388b67143644db4a412e4c888f8c796068daff29f63615f0b6a3a04262a", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/AndroidArenaAgent/AndroidArena/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:24:52Z", "role": "primary", "sha256": "cfa9d272c85398f54c18d6e7445d680009e8c16d128962235d899592d3094c2f", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/observation.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:24:53Z", "role": "primary", "sha256": "d25f2edf9202f59615341b62fc6046ac23a6f4cb072a6038ac562506cbdc891c", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/actions.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "257c42e980a0b14bd74bfcc381a0ecbef2b538c0f1cc1a94d22925efe354fb8e", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/run_evaluator.py"}], "id": "github:androidarenaagent/androidarena", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AndroidArenaAgent/AndroidArena"}], "name": "AndroidArena", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/README.md", "target_id": "arxiv:2402.06596", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Android environment for evaluating language-model-driven agents.", "topics": ["evaluation", "mobile-navigation"], "updated_at": null, "url": "https://github.com/AndroidArenaAgent/AndroidArena"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://api.github.com/repos/anikasomaia/local-activity-recorder/git/trees/f7ec5b6ecf0ba1bfe6362d7f2bfb0f1af5ac6bc7?recursive=1"], "reason": "No license declaration was found in the inspected README/C# files, and the complete pinned tree has no LICENSE, LICENCE or COPYING-named file. Public source access does not establish reuse permission."}}, "repository_revision": "967cea087bf6496b3b82743fb96770fa29f8e203"}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:57:15.716415+00:00", "role": "primary", "sha256": "440e0e43026ba7a20eaca17507cec33bd697d8b1eed6fb056be34136b3f4f76f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/anikasomaia/local-activity-recorder/967cea087bf6496b3b82743fb96770fa29f8e203/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:04:41.685729+00:00", "role": "primary", "sha256": "90c30399c8aa5031f82164978bc12eae3bebba172758d20877a87fec53566c12", "supports": ["identity"], "url": "https://api.github.com/search/repositories?q=topic%3Acomputer-use-agents+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:54:07.061516+00:00", "role": "primary", "sha256": "f5b501fa132163d749c6695d81c76d45fdff2888ea890e31cbaa9bc4172fed3b", "supports": ["identity", "details.repository_revision", "summary", "environments", "topics", "notes"], "url": "https://api.github.com/repos/anikasomaia/local-activity-recorder/commits/master"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:56:33.833893+00:00", "role": "primary", "sha256": "ee9e5899d1f150982eb05ca1f8794406cef0ffc80eb8223811b504ff972370c1", "supports": ["identity", "details.repository_revision", "summary", "environments", "notes", "licenses.code"], "url": "https://api.github.com/repos/anikasomaia/local-activity-recorder/git/trees/f7ec5b6ecf0ba1bfe6362d7f2bfb0f1af5ac6bc7?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:05:00.168119+00:00", "role": "primary", "sha256": "ce3022b4016d140eac554e392dbe42f54ffb3e8098b03e2bf72e92ff49c5d084", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/anikasomaia/local-activity-recorder/967cea087bf6496b3b82743fb96770fa29f8e203/Program.cs"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:05:03.421222+00:00", "role": "primary", "sha256": "e70e2e68b5dd586e5ad0fb0b12c8b937b4614bd5237bac61d03d22144c4751f4", "supports": ["notes"], "url": "https://raw.githubusercontent.com/anikasomaia/local-activity-recorder/967cea087bf6496b3b82743fb96770fa29f8e203/DataUploader.cs"}], "id": "github:anikasomaia/local-activity-recorder", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-20T11:57:15.716415+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/anikasomaia/local-activity-recorder"}], "name": "Activity Recorder", "notes": "Recording is started and stopped from the tray; screenshots use a five-second timer. The manual Upload Data action uses a built-in third-party demonstration endpoint unless changed in code, sending recorded files as base64 JSON. Successful uploads move files into uploaded subfolders rather than delete them; no automatic expiry policy was found in the inspected capture/upload path. Sensitive-application filters are listed as future work. The README's .NET version requirements are inconsistent, and long-session behavior was not established. No license declaration was found in the inspected files or complete tree's license-file names, so code permission remains unknown. No activity was captured and no uploader or executable was run.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Windows tray utility that records primary-display screenshots, keyboard key codes and mouse events as local JPEG and JSON files for training-data collection.", "topics": ["recording", "data-generation"], "updated_at": null, "url": "https://github.com/anikasomaia/local-activity-recorder"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["list of UI element boxes [x1,y1,x2,y2] and text descriptions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md", "population": "English annotation subset reported by README", "qualifier": "exact", "unit": "page-annotation entries", "value": 1000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md", "population": "Chinese annotation subset reported by README", "qualifier": "exact", "unit": "page-annotation entries", "value": 1000}], "evaluation": "Metrics use greedy one-to-one matching in descending IoU order, then micro precision/recall/F1 and lowercased difflib.SequenceMatcher description similarity on matched pairs. Actual metrics.py accepts IoU >= 0.5 at its default threshold; README prose says > 0.5. Predictions align to annotations by image_path; callers must supply complete prediction coverage because missing predictions are not independently enumerated by evaluate_dataset.", "evidence_urls": ["https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md", "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/src/metrics.py", "https://api.github.com/repos/ant-research/GUIHalluBench/git/trees/master?recursive=1", "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/LEGAL.md", "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/.gitattributes"], "focus": ["GUI page-element detection", "element-description consistency"], "notes": "Each annotation entry represents a page, not one UI element. README’s illustrative num_samples=4796 output is not adopted as a corpus count. Public tree contains both annotation JSON files and an images.tar.gz Git-LFS pointer; payload availability/integrity was not tested by downloading images. LEGAL.md only governs comment-language precedence, not a code/data license grant; both slots remain unknown. No benchmark run was performed.", "observations": ["page screenshot"], "setting": "offline"}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:51Z", "role": "primary", "sha256": "51df9dbcf52f46e4774207f2e6ad84dc8a6d117cdab6f77a38e9897ded1d0963", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://github.com/ant-research/GUIHalluBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "d9c476d6960a4ef99e20d4c3b5d35a63f3ae5e61fc20e9dea00b11e64b74c9e9", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "f36707dfcd3e83fc9d7f57b50d0c27007801e5029cc1ac255e000560ecf778fc", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/src/metrics.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:07Z", "role": "primary", "sha256": "c0896bc90d515b7bd8fe86b062aeb199bd47570e6eca8474986f5b1ef1d98dbd", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/ant-research/GUIHalluBench/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "adba50d9794b9ef3f7ec8cbc680f7f1fa3fbf9df0ac8d1f9b9ccab6d941bc11b", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/LEGAL.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:34:56Z", "role": "primary", "sha256": "8f35fa75be42a7a53e4f96de1b62be84962d2cf1beb8b23dfcfaecfaea4a6b90", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/.gitattributes"}], "id": "github:ant-research/guihallubench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ant-research/GUIHalluBench"}], "name": "GUIHalluBench", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Measures hallucination in UI-element parsing by matching the boxes and descriptions a model lists for 1,000 English and 1,000 Chinese screenshots against human annotations.", "topics": ["evaluation", "robustness"], "updated_at": null, "url": "https://github.com/ant-research/GUIHalluBench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "0d4b28c993ccb977c7910ff3fecb5de29db364a7a7ad55523c8e052c9bae2566", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "b02ec75f69ab8bbebe0d59eeccdcdcacf11a0d6b00bdfe803c692a9d587301e2", "supports": ["identity"], "url": "https://api.github.com/repos/ApGa/Go-Browse"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "cd3a40a596ad9ead856973bb31e5b7f7e88a2f60969799c01a2a8a6d09fe3421", "supports": ["release_status"], "url": "https://api.github.com/repos/ApGa/Go-Browse/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "e7107f1e1b3a0c65cbbae6c496d9e71b570025bdddebb3d402a6080dd916bcba", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "e7107f1e1b3a0c65cbbae6c496d9e71b570025bdddebb3d402a6080dd916bcba", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/LICENSE"}], "id": "github:apga/go-browse", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:44:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T02:45:20Z", "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ApGa/Go-Browse"}], "name": "Go-Browse", "notes": "The release contains exploration agents, data-processing scripts, fine-tuning and WebArena evaluation entry points under MIT. Checkpoints and processed datasets are separate Hub resources with separate license status. The arXiv badge image has an unrelated number; the linked paper and citation consistently identify 2506.03533.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md", "target_id": "arxiv:2506.03533", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Structured website-exploration and data-processing code that revisits discovered pages, proposes tasks and collects trajectories for web-agent fine-tuning.", "topics": ["web-navigation", "data-generation", "imitation-learning"], "updated_at": null, "url": "https://github.com/ApGa/Go-Browse"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "ecd5b8033cbc17a2ce5aebf1b35260b77c4906aea7c549a0b86e5d55b9ebe135", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/appium/appium/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "5cb791a45cafc4bdaef57f3ba49fb7691a98c756e8148467f9e241295728490f", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/appium/appium/HEAD/LICENSE"}], "id": "github:appium/appium", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/appium/appium/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/appium/appium"}], "name": "Appium", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Open-source WebDriver-based automation server for mobile, desktop and IoT apps, extended by per-platform drivers for iOS, Android, macOS and Windows, plugins, and clients in Java, Python, Ruby and C#.", "topics": ["automation"], "updated_at": null, "url": "https://github.com/appium/appium"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:52Z", "role": "primary", "sha256": "1973f702c08ca9637ce03bc28fc34188be05826e9d43f9992b08e804f474b1a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/apple-aiml-research/ml-ferret/main/ferretui/README.md"}], "id": "github:apple/ml-ferret", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:40:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/apple/ml-ferret"}], "name": "Ferret-UI", "notes": "The repository has moved from apple/ml-ferret to apple-aiml-research/ml-ferret; evidence uses the dedicated ferretui component.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/apple-aiml-research/ml-ferret/main/ferretui/README.md", "target_id": "arxiv:2404.05719", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screen understanding research within the Ferret model family.", "topics": ["screen-understanding", "grounding"], "updated_at": null, "url": "https://github.com/apple/ml-ferret"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a2a3c5ad115e4e2109b56686c9e0b073bcfec6d9f052a58bc92683af44933480", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"}], "id": "github:ariaui/aria-ui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AriaUI/Aria-UI"}], "name": "Aria-UI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "target_id": "arxiv:2412.16256", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mixture-of-experts GUI grounding model with 3.9B activated parameters that can use text or interleaved image history, with released training data and an AndroidWorld agent.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/AriaUI/Aria-UI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "9e7a23adb62bd05d0fd45e5c95a57fcfc3cc4d19ab89d6801c5bf1e23d3c678c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/asweigart/pyautogui/HEAD/README.md"}], "id": "github:asweigart/pyautogui", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/asweigart/pyautogui"}], "name": "PyAutoGUI", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Cross-platform Python module for scripted mouse and keyboard control, message boxes and screenshots with on-screen image search, working on the primary monitor under Windows, macOS and Linux.", "topics": ["automation"], "updated_at": null, "url": "https://github.com/asweigart/pyautogui"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "fe475d3e92c42807b792c542bc5986bf74caa7d166aea860d368fbe776342a82", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/BAAI-Agents/Cradle/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "2964861ad96d977da95d891f62d8f5d60873f4a36da0b51f052d9e98ef97d02f", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/BAAI-Agents/Cradle/HEAD/LICENSE"}], "id": "github:baai-agents/cradle", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/BAAI-Agents/Cradle/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/BAAI-Agents/Cradle"}], "name": "Cradle", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/BAAI-Agents/Cradle/HEAD/README.md", "target_id": "arxiv:2403.03186", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Framework in which foundation models control games and software from screenshots alone, acting through keyboard and mouse, shown on RDR2, Stardew Valley, Cities: Skylines and Dealer's Life 2.", "topics": ["planning", "memory"], "updated_at": null, "url": "https://github.com/BAAI-Agents/Cradle"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/README.md", "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/LICENSE"], "reason": "The retained README restricts SheetCopilot and its dataset to noncommercial use, while the retained repository LICENSE contains the GNU GPL version 3 text. These declarations do not establish one unambiguous code license."}}}, "environments": ["desktop", "windows", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:56Z", "role": "primary", "sha256": "beb8c811bf2d070845aa6fcc7e94ffd43ca4f9f60ce42c72d958598790fbdffc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.code", "notes"], "url": "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:47Z", "role": "primary", "sha256": "3972dc9744f6499f0f9b2dbf76696f2ae7ad8af9b23dde66d6af86c9dfb36986", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:47Z", "role": "primary", "sha256": "3972dc9744f6499f0f9b2dbf76696f2ae7ad8af9b23dde66d6af86c9dfb36986", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/LICENSE"}], "id": "github:bravegroup/sheetcopilot", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-18T10:59:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/BraveGroup/SheetCopilot"}], "name": "SheetCopilot", "notes": "The original agent and the later Ubuntu evaluator have different roles; the repository does not establish that the original agent runs on Ubuntu. Code licensing is unresolved because the retained README and LICENSE contain conflicting declarations.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/README.md", "target_id": "arxiv:2305.19308", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Spreadsheet automation agent for Windows/Excel, with benchmark assets and a separately released headless Ubuntu evaluation backend.", "topics": ["workflow", "tool-integration", "evaluation"], "updated_at": null, "url": "https://github.com/BraveGroup/SheetCopilot"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "fcca30a036d122e34c943740c29d8cb91642f0c615050a8d54ec891c0eee56d2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/browser-use/browser-use/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "135c57671b0ee957669b05a0ca00cc06f645496d351e2e732832c6e276da9a5e", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/browser-use/browser-use/HEAD/LICENSE"}], "id": "github:browser-use/browser-use", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/browser-use/browser-use/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/browser-use/browser-use"}], "name": "Browser Use", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/browser-use/browser-use/HEAD/README.md", "target_id": "organization:browser-use", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Open-source browser agent library that runs locally with any LLM against a local or cloud browser, and also serves as a CLI for coding agents and as a driver for Claude's browser toolset.", "topics": ["web-navigation"], "updated_at": null, "url": "https://github.com/browser-use/browser-use"}
{"aliases": [], "authors": [], "details": {"availability": "Published source; official release metadata inspected", "release_tag": "v3.0.0"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "5047bcf2f0e92f0d34a57ac7d83166903a16a9909c700c0647b8a48363bd94f9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/browser-use/web-ui/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "7fb0738bc9b837f7b3e645c406e04d5c7c4d0860ea636a96e4e9d84d9510dd82", "supports": ["identity"], "url": "https://api.github.com/repos/browser-use/web-ui"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "fcbf1775e3513625103d389c737e98e939e17140baac9bc8000b582ab2729a46", "supports": ["release_status"], "url": "https://api.github.com/repos/browser-use/web-ui/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:00Z", "role": "primary", "sha256": "f87ee752c38653c447eed4a9abd83e71dedd40298758bcb9aa993b731362c453", "supports": ["release_status"], "url": "https://api.github.com/repos/browser-use/web-ui/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:01Z", "role": "primary", "sha256": "612d1382ec1a426eaf3672aafb13bafcfe91b2c90a0fe339822ffc9c895003ca", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/browser-use/web-ui/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:01:01Z", "role": "primary", "sha256": "612d1382ec1a426eaf3672aafb13bafcfe91b2c90a0fe339822ffc9c895003ca", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/browser-use/web-ui/HEAD/LICENSE"}], "id": "github:browser-use/web-ui", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T04:01:01Z", "url": "https://raw.githubusercontent.com/browser-use/web-ui/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/browser-use/web-ui"}], "name": "Browser Use WebUI", "notes": "The repository publishes UI source and release v3.0.0 under MIT. It is a distinct user interface built on the existing Browser Use framework, not a second framework identity. The UI supports several external model providers whose terms are separate; no dependencies were installed.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A Gradio interface for configuring and running Browser Use agents, including persistent browser sessions and interaction recordings.", "topics": ["web-navigation", "tool-integration", "recording"], "updated_at": null, "url": "https://github.com/browser-use/web-ui"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "fb2039ee569e37cf1d652ca26c0a1951a318d4ac3240018b958985c0aec91007", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/browserbase/stagehand/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "fac7e7cd1cfc504054ee1fc53c38e0f3099d7357917e85a73792f6511fe0ea96", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/browserbase/stagehand/HEAD/LICENSE"}], "id": "github:browserbase/stagehand", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/browserbase/stagehand/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/browserbase/stagehand"}], "name": "Stagehand", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/browserbase/stagehand/HEAD/README.md", "target_id": "organization:browserbase", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Browser automation framework combining code and natural-language actions.", "topics": ["web-navigation"], "updated_at": null, "url": "https://github.com/browserbase/stagehand"}
{"aliases": [], "authors": [], "details": {"availability": "Published source; official release metadata inspected", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/LICENSE", "https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/README.md"], "reason": "The repository explicitly identifies its code as AGPL-3.0; this declaration does not cover every incorporated third-party component."}}, "release_tag": "browserclaw/v0.51.1.2"}, "environments": ["web", "desktop", "windows", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "bfe2979ae0e99985154042090f483e172cca229fe408f8f9b47102662beec983", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.code"], "url": "https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "2b8c90b9fb2edffccf7406952465015df87622331921fe0f25a44fc9f5a9b86b", "supports": ["identity"], "url": "https://api.github.com/repos/browseros-ai/BrowserOS"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "1f43b364f532a8bdd798df5dd86d3b73f8f545ef73a66e9b3b4da59ea9d0faa1", "supports": ["release_status"], "url": "https://api.github.com/repos/browseros-ai/BrowserOS/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:59Z", "role": "primary", "sha256": "87bea16220429a93fefc2b2a57452fd67b2618784025db18eb1f899c0396098f", "supports": ["release_status"], "url": "https://api.github.com/repos/browseros-ai/BrowserOS/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:59Z", "role": "primary", "sha256": "8486a10c4393cee1c25392769ddd3b2d6c242d6ec7928e1414efff7dfb2f07ef", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/LICENSE"}], "id": "github:browseros-ai/browseros", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": {"basis": "Repository LICENSE and explicit README declaration; incorporated Chromium and other components retain their own terms.", "declared": "AGPL-3.0", "retrieved_at": "2026-10-17T04:00:59Z", "url": "https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/browseros-ai/BrowserOS"}], "name": "BrowserOS / BrowserOS neo", "notes": "The one repository includes BrowserOS neo and BrowserOS; they are retained as one source-bundle resource. Official release metadata lists desktop installers, but none were downloaded. The repository declares AGPL-3.0 and retains separate Chromium-related licensing. Publisher privacy and performance claims were not independently verified.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Chromium-based browser implementations for external agent control and built-in assistant workflows, with session visibility and replay facilities.", "topics": ["web-navigation", "tool-integration", "recording"], "updated_at": null, "url": "https://github.com/browseros-ai/BrowserOS"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "796cc24df9874c71408d32d12135b6b04389c455134c6c722ca34711288beb40", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/bytebot-ai/bytebot/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/bytebot-ai/bytebot/HEAD/LICENSE"}], "id": "github:bytebot-ai/bytebot", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/bytebot-ai/bytebot/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bytebot-ai/bytebot"}], "name": "Bytebot", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Open-source desktop agent that works inside its own Ubuntu Linux virtual desktop, with a web task interface, REST APIs for tasks and direct control, a takeover mode and password-manager logins.", "topics": ["workflow"], "updated_at": null, "url": "https://github.com/bytebot-ai/bytebot"}
{"aliases": ["github:bytedance/agent-r", "bytedance/Agent-R", "https://github.com/bytedance/Agent-R"], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:20Z", "role": "primary", "sha256": "8471a0761a94b1af84099a80efb6ac53f17e34491a196e82ca068a5113c0b8be", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ByteDance-Seed/Agent-R/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:19Z", "role": "primary", "sha256": "3fff7137bc4f39a3a757f3b820892fb330449044ef2721e9f9d7719d23ff2b98", "supports": ["identity", "availability"], "url": "https://api.github.com/repos/ByteDance-Seed/Agent-R"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "e3c70b04094e8e77e249065b2b500ac197fc6fafee6e3275d78289b3790bf7ed", "supports": ["release_status"], "url": "https://api.github.com/repos/ByteDance-Seed/Agent-R/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:22Z", "role": "primary", "sha256": "283ea6cc2997a1a70da0049e09adf9317bb60ca1b51279b65196b83a69e1996b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ByteDance-Seed/Agent-R/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:05:22Z", "role": "primary", "sha256": "283ea6cc2997a1a70da0049e09adf9317bb60ca1b51279b65196b83a69e1996b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ByteDance-Seed/Agent-R/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T18:05:46Z", "role": "primary", "sha256": "3fff7137bc4f39a3a757f3b820892fb330449044ef2721e9f9d7719d23ff2b98", "supports": ["aliases"], "url": "https://api.github.com/repos/bytedance/Agent-R"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T18:05:46Z", "role": "primary", "sha256": "3fff7137bc4f39a3a757f3b820892fb330449044ef2721e9f9d7719d23ff2b98", "supports": ["aliases"], "url": "https://api.github.com/repos/ByteDance-Seed/Agent-R"}], "id": "github:bytedance-seed/agent-r", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:03:20Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T02:05:22Z", "url": "https://raw.githubusercontent.com/ByteDance-Seed/Agent-R/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ByteDance-Seed/Agent-R"}], "name": "Agent-R", "notes": "The canonical repository is ByteDance-Seed/Agent-R and is archived but readable. MCTS collection, revision and evaluation scripts are released under Apache-2.0. It also supports non-GUI text environments; this entry covers its WebShop interaction component. No separate weight release was verified.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ByteDance-Seed/Agent-R/HEAD/README.md", "target_id": "arxiv:2501.11425", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "Trajectory collection and revision utilities that use tree search and model-guided critique to train agents to recover from mistakes, including WebShop tasks.", "topics": ["self-reflection", "data-generation", "web-navigation"], "updated_at": null, "url": "https://github.com/ByteDance-Seed/Agent-R"}
{"aliases": [], "authors": [], "details": {"availability": "Public desktop data-generation source; separate training-data and trained-model releases not verified.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/LICENSE"], "reason": "Repository code license; no scope transfer to generated trajectories, pretrained detector weights, or GUI application content."}}, "repository_revision": "78a65d003d6f48802fac879508b62f1c89125892"}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:15Z", "role": "primary", "sha256": "63708cb2500aefa36916f3fb171b595efaabf614f17b925cfe8b5c8cbf72be10", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:15Z", "role": "primary", "sha256": "34addf6eed41e1e29e51aab8819c5ac316b5699a8c11c4ce29a32c8e38557de8", "supports": ["identity"], "url": "https://api.github.com/repos/bytedance/GUI-ReWalk"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:07Z", "role": "primary", "sha256": "a218cb6f8a2c2a243e8207785c98255bf717c00f177521368705d8652e229d7d", "supports": ["release_status"], "url": "https://api.github.com/repos/bytedance/GUI-ReWalk/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:08Z", "role": "primary", "sha256": "43070e2d4e532684de521b885f385d0841030efa2b1a20bafb76133a5e1379c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/LICENSE"}], "id": "github:bytedance/gui-rewalk", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T09:12:15Z", "licenses": {"code": {"basis": "Repository code license; no scope transfer to generated trajectories, pretrained detector weights, or GUI application content.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T09:14:08Z", "url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bytedance/GUI-ReWalk"}], "name": "GUI-ReWalk", "notes": "The inspected source tree contains the random walker, desktop environment integration and run_rewalk.sh. The documented runnable setup uses OSWorld and OmniParser. The README also discusses mobile data and a GUI-ReWalk-7B model, but no separately published corpus or model checkpoint was verified; those are not counted as released assets. Apache-2.0 is scoped to the implementation.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/README.md", "target_id": "github:xlang-ai/osworld", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/README.md", "target_id": "github:microsoft/omniparser", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/README.md", "target_id": "arxiv:2509.15738", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A desktop trajectory-generation framework that moves from stochastic GUI exploration to inferred goals and multi-stage task completion, with retrospective reasoning annotations.", "topics": ["data-generation", "planning", "cross-app"], "updated_at": null, "url": "https://github.com/bytedance/GUI-ReWalk"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "0c4feb17b657c0c9a71d7e0c5e2142abbd19bb5cff22e6cd3b6b082644c98ee7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/LICENSE"}], "id": "github:bytedance/ui-tars", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bytedance/UI-TARS"}], "name": "UI-TARS", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/README.md", "target_id": "arxiv:2501.12326", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Native GUI agent models; UI-TARS-1.5 adds reinforcement-learned reasoning before acting and an open 7B checkpoint, and UI-TARS-2 extends to games, code and tool use.", "topics": ["grounding", "planning"], "updated_at": null, "url": "https://github.com/bytedance/UI-TARS"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "0f05aa4f447d43439f6b3f4ab179cbe1e59d170eaf6c0684fa9c2eb0872e5985", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/bytedance/UI-TARS-desktop/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/bytedance/UI-TARS-desktop/HEAD/LICENSE"}], "id": "github:bytedance/ui-tars-desktop", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/bytedance/UI-TARS-desktop/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bytedance/UI-TARS-desktop"}], "name": "UI-TARS Desktop", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Desktop app for a local GUI agent driven by UI-TARS and Seed-1.5-VL/1.6 models, offering local and remote operators for computers and browsers; the repository also ships the Agent TARS CLI and web UI.", "topics": ["workflow"], "updated_at": null, "url": "https://github.com/bytedance/UI-TARS-desktop"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "dc60f683b0f44ccee4e54e5cbe1552ee3454e1ec4ed5c04635fedb83a325f51f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/bz-lab/AUITestAgent/HEAD/README.md"}], "id": "github:bz-lab/auitestagent", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bz-lab/AUITestAgent"}], "name": "AUITestAgent", "notes": "Its README describes mobile apps without naming an operating system.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/bz-lab/AUITestAgent/HEAD/README.md", "target_id": "arxiv:2407.09018", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Generates mobile GUI interactions from natural-language test requirements and checks application behavior against those requirements.", "topics": ["automation", "evaluation"], "updated_at": null, "url": "https://github.com/bz-lab/AUITestAgent"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/caap-agent/caap-agent/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate. The inspected text does not establish -only versus -or-later for the project."}}, "repository_tree_sha": "9d65c0f30e3c119db533dd832863b4f3f7b74b18", "repository_url": "https://github.com/caap-agent/caap-agent"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "481db1288ffc769f48b19c83f0db4fb40a64dc1b05859f776b77f0774bfcdb90", "supports": ["name", "kind", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/caap-agent/caap-agent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:02Z", "role": "primary", "sha256": "8d518cad746591e82b54b63b79d0a1d47531a75f02c8baf2b560877fb03918ba", "supports": ["identity"], "url": "https://api.github.com/repos/caap-agent/caap-agent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:24Z", "role": "primary", "sha256": "c511d94ce58e73a87c9647169b1199c0ed0029e6b1df271ad1fcd3ca86f3458b", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/caap-agent/caap-agent/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["summary", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:20Z", "role": "primary", "sha256": "5e6fa1bfd473724e67366c42761db27fe091ee00fdcfbe599b9dd5d5a9311d36", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/caap-agent/caap-agent/HEAD/LICENSE"}], "id": "github:caap-agent/caap-agent", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-18T04:27:14Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate. The inspected text does not establish -only versus -or-later for the project.", "declared": "AGPL-3.0", "retrieved_at": "2026-10-18T04:35:20Z", "url": "https://raw.githubusercontent.com/caap-agent/caap-agent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/caap-agent/caap-agent"}], "name": "CAAP Agent", "notes": "Code is released for the two named browser benchmarks. Visual-observer checkpoints are linked separately through Google Drive and were not downloaded or access-tested. Root LICENSE begins with GNU AGPL v3 and includes third-party notices; this code declaration is not transferred to bundled/generated data or external weights.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50", "target_id": "arxiv:2406.06947", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screenshot-driven agent implementation for MiniWoB++ and WebShop, using a visual observer and context-aware prompts to issue mouse and keyboard actions.", "topics": ["planning", "screen-understanding", "web-navigation"], "updated_at": null, "url": "https://github.com/caap-agent/caap-agent"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["Python-defined environment actions"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50", "population": "CRAB Benchmark-v0 (paper)", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2407.01511", "unit": "tasks", "value": 120}], "evaluation": "Graph-based evaluators measure progress through task requirements.", "evidence_urls": ["https://raw.githubusercontent.com/camel-ai/crab/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"], "focus": ["cross-environment tasks"], "notes": "CRAB is also an extensible framework. The 120-task count belongs to Benchmark-v0, not every environment that can be created with the framework.", "observations": [], "setting": "interactive"}}, "environments": ["desktop", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "7b894c5474ba78294427cfd4ef0eea8cb22d76dfa0c2e8c7f01e31ad927064e9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/camel-ai/crab/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "github:camel-ai/crab", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/camel-ai/crab"}], "name": "CRAB", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/camel-ai/crab/HEAD/README.md", "target_id": "arxiv:2407.01511", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Cross-environment benchmark framework for multimodal agents.", "topics": ["evaluation", "cross-app"], "updated_at": null, "url": "https://github.com/camel-ai/crab"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["READ text answer", "ACT selector, text/label, accessible role/name, DOM index, Stagehand reference or screenshot point, according to the representation"], "counts": [], "evaluation": "READ uses normalized exact match against an answer or allowed alias. ACT resolves predicted identifiers to ground-truth elements, or checks screenshot points inside the target box. Representation-token accounting subtracts fixed-prompt tokens; screenshots use billed input usage minus the fixed prompt. These are answer/target correctness and token metrics, not completed-task success.", "evidence_urls": ["https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/README.md", "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/METHODOLOGY.md", "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/src/wpbench/scoring.py"], "focus": ["Browser-representation token use", "Factual READ accuracy", "ACT target-element selection"], "notes": "Fixed local-page observations; the reported run uses claude-sonnet-5 and a shared prompt, without scrolling or an agent loop. Browser Use disables its viewport cutoff and sends untruncated text; Stagehand uses a separate matched browser load. Short-string representability labels are approximate. Offline does not mean API-free. Third-party snapshot redistribution remains unresolved; results were not reproduced.", "observations": ["Top-document HTML", "Markdown converted from HTML", "Playwright accessibility snapshot", "Browser Use indexed DOM", "Stagehand hybrid snapshot", "Viewport screenshot"], "setting": "offline"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/LICENSE"], "reason": "The inspected root LICENSE contains complete MIT terms for repository software. The methodology explicitly leaves redistribution permission for captured third-party practice-site pages unresolved; code MIT is not a blanket snapshot/data license."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:30.874397Z", "role": "primary", "sha256": "14775f6e3daba69d4684380fd33bf717bdc3093b65be796eb863bbd27d0c0bab", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "environments", "details.benchmark_protocol.focus", "details.benchmark_protocol.setting", "details.benchmark_protocol.observations", "details.benchmark_protocol.actions", "details.benchmark_protocol.notes"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:05:35.063004+00:00", "role": "primary", "sha256": "3e22e1cbfa47bcf98f6f70a5f6f162a31a509a9f2091a84679bb5231977e29d9", "supports": ["id", "url"], "url": "https://api.github.com/search/repositories?q=topic%3Aweb-agents+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:50:54.545491Z", "role": "primary", "sha256": "fd30e94e78e7167a49579922259fde1c7d7f8259e9f00afba15faf3a71966a94", "supports": ["id", "url", "notes"], "url": "https://github.com/charlypoly/web-context-bench-python"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:30.131541Z", "role": "primary", "sha256": "1c3c0cbf9f9d9f7cd9e7a2f027ecb217fd5c16fa157916281a15bb7f403f6f1c", "supports": ["summary", "environments", "topics", "scope", "notes", "licenses.data", "details.benchmark_protocol.focus", "details.benchmark_protocol.setting", "details.benchmark_protocol.observations", "details.benchmark_protocol.actions", "details.benchmark_protocol.evaluation", "details.benchmark_protocol.notes"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/METHODOLOGY.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:28.567861Z", "role": "primary", "sha256": "8cada5d71b5a61e7aff2e1ac2b951c266eb2db7c082c06c5a13aa127d42b3fc5", "supports": ["summary", "topics", "scope", "notes", "details.benchmark_protocol.evaluation", "details.benchmark_protocol.notes"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/src/wpbench/scoring.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:41.499735Z", "role": "primary", "sha256": "9b64ef672f5091b6074d02502f6d3c161d94308a3d9d142d7437dcddf611bbc4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/LICENSE"}], "id": "github:charlypoly/web-context-bench-python", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-20T11:51:30.874397Z", "licenses": {"code": {"basis": "Complete MIT license text in the repository root at the inspected commit; repository code only.", "declared": "MIT", "retrieved_at": "2026-10-20T11:51:41.499735Z", "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/charlypoly/web-context-bench-python"}], "name": "Web Context Bench (Python)", "notes": "This is a single-snapshot representation evaluation on local fixtures and captured pages, not a live multi-step agent benchmark. Its shared-prompt, one-model setup differs from complete Browser Use or Stagehand agents. The methodology notes unresolved redistribution licensing for third-party practice-site snapshots; MIT is recorded for repository code only. Selected scoring code was read, but no models, browser tasks or claimed results were run or reproduced.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Local-page benchmark comparing six browser representations on token use, factual reading and target-element selection, with deterministic answer scoring.", "topics": ["evaluation", "grounding", "efficiency", "screen-understanding"], "updated_at": null, "url": "https://github.com/charlypoly/web-context-bench-python"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/LICENSE"], "reason": "The inspected root LICENSE is complete MIT text for repository code. The local data README says same as parent repository, but the linked Hub dataset/model cards and artifacts were not inspected; their licenses are not inferred."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:34.292633Z", "role": "primary", "sha256": "7cda5207d8882bc8dedaefb93d2ccf54b310354917283b8f4890fffdc5764958", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "environments", "relations"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:04:53.470789+00:00", "role": "primary", "sha256": "3d76d934e98641746085772b98ee17059749f980a616435498a7ad483bd4cec1", "supports": ["id", "url"], "url": "https://api.github.com/search/repositories?q=topic%3Agui-grounding+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:50:58.795155Z", "role": "primary", "sha256": "08b89c5e6143afa4c0cdde95f8ec016d9033ee004f7631fadd4457144c2b87c9", "supports": ["id", "url", "notes"], "url": "https://github.com/ChenghengLi/WinDOM"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:41.278660Z", "role": "primary", "sha256": "09488140d35408cc11bb9c12b1bb435b21814cf6525937a8309475c9326c4004", "supports": ["environments", "notes", "licenses.data"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/data/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:45.091369Z", "role": "primary", "sha256": "0ee33b1e416ea4c5d8834688e6f42fbd15ef493f1b82398d3bb985838cbdffb9", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/src/windom/training/sfd_4b.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:48.511963Z", "role": "primary", "sha256": "c481b2f91f2d82bf801632da6ab6cce01bc42c780104d743156c272942d2cef3", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/src/windom/evaluation/screenspot.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:44.940106Z", "role": "primary", "sha256": "d5a37ff86df9b231440ab1bcf3ecf77304a0a53e772182cdb622a3ab5d043447", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/LICENSE"}], "id": "github:chenghengli/windom", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-20T11:51:34.292633Z", "licenses": {"code": {"basis": "Complete MIT license text in the repository root at the inspected commit; repository code only.", "declared": "MIT", "retrieved_at": "2026-10-20T11:51:44.940106Z", "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ChenghengLi/WinDOM"}], "name": "WinDOM code", "notes": "The Windows-style environment is win11React in Playwright, not native Windows control. The repository documents its predecessor trajectory pipeline but explicitly omits the final corpus harvester. The inspected ScreenSpot evaluator defaults to a privileged ground-truth-center demonstration; --no-demo and --official-prompt select other modes, so evaluation results require their prompt and coordinate settings. Linked datasets, weights and reported scores were not independently verified; MIT applies to inspected repository code.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/README.md", "target_id": "arxiv:2606.25964", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI-grounding training and evaluation code with same-family distillation, GRPO recipes and point-in-box scoring, using a Windows-style web environment.", "topics": ["grounding", "imitation-learning", "reinforcement-learning", "evaluation"], "updated_at": null, "url": "https://github.com/ChenghengLi/WinDOM"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "0759e1212025ac487c1646877f97526d35db1a62337192800acaff7684471087", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ChenWu98/agent-attack/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "github:chenwu98/agent-attack", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ChenWu98/agent-attack"}], "name": "Adversarial Web Agents", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50", "target_id": "arxiv:2406.12814", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Evaluation resources for adversarial attacks on multimodal web agents.", "topics": ["safety-security", "robustness"], "updated_at": null, "url": "https://github.com/ChenWu98/agent-attack"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "93defc2b31d0f8e6144e77cda4a74abeafd5253ae18a626bfc625f429050770a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/chessbean1/GUI-Robust/HEAD/README.md"}], "id": "github:chessbean1/gui-robust", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/chessbean1/GUI-Robust"}], "name": "GUI-Robust", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/chessbean1/GUI-Robust/HEAD/README.md", "target_id": "arxiv:2506.14477", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Evaluation scripts and linked data for comparing GUI action predictions and task execution under ordinary and anomalous interface conditions.", "topics": ["evaluation", "robustness"], "updated_at": null, "url": "https://github.com/chessbean1/GUI-Robust"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "40a5e9ad3a2169527a9330825bd6e07ac1669293e21c209ed41036987bbf0a8b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/china-qijizhifeng/Work2Skill/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "70387bc43f567b2ed971c1a789719c8c3567b0d073407c38b5dc6af1eb78012a", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/china-qijizhifeng/Work2Skill/HEAD/LICENSE"}], "id": "github:china-qijizhifeng/work2skill", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/china-qijizhifeng/Work2Skill/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/china-qijizhifeng/Work2Skill"}], "name": "Work2Skill", "notes": "The README reports that evaluation integration and live cross-platform acceptance testing remain pending.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/china-qijizhifeng/Work2Skill/HEAD/README.md", "target_id": "arxiv:2610.12984", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Collects human GUI demonstrations and inspects multimodal evidence to derive reusable agent workflow skills.", "topics": ["recording", "workflow", "data-generation"], "updated_at": null, "url": "https://github.com/china-qijizhifeng/Work2Skill"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "059a0349daaef6bda98eae6d9384ed349ac94b8033b18be817c317a01b472024", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ChromeDevTools/chrome-devtools-mcp/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "58d1e17ffe5109a7ae296caafcadfdbe6a7d176f0bc4ab01e12a689b0499d8bd", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ChromeDevTools/chrome-devtools-mcp/HEAD/LICENSE"}], "id": "github:chromedevtools/chrome-devtools-mcp", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/ChromeDevTools/chrome-devtools-mcp/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ChromeDevTools/chrome-devtools-mcp"}], "name": "Chrome DevTools MCP", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "MCP tools for browser control and inspection through Chrome DevTools.", "topics": ["automation", "tool-integration"], "updated_at": null, "url": "https://github.com/ChromeDevTools/chrome-devtools-mcp"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "a2df62f9b4359f7839c7d6c6917ff654f9eadeafa92f96311495b71e37abd9ba", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/convergence-ai/webgames/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "0e03b68c7acb0b340e47f9000ac3808950e2ac8c4cd05960437a9e46e3c09a62", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/convergence-ai/webgames/HEAD/LICENSE"}], "id": "github:convergence-ai/webgames", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/convergence-ai/webgames/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/convergence-ai/webgames"}], "name": "WebGames", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "150 client-side web challenges in five capability domains, built to be easy for humans but hard for browsing agents, each revealing a password when solved.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/convergence-ai/webgames"}
{"aliases": ["github:cooelf/auto-ui", "cooelf/Auto-UI", "https://github.com/cooelf/Auto-UI"], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "13f045fa84dc0ad68dcceea95fc7e0e34a45bcda866d589132351559b87a8e30", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/cooelf/Auto-GUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/cooelf/Auto-GUI/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:13:59Z", "role": "primary", "sha256": "569f3c779e6a0b8210a764c9096e49e8a5199c107a64f7fcfb7a011257bf2a12", "supports": ["aliases"], "url": "https://api.github.com/repos/cooelf/Auto-UI"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:00Z", "role": "primary", "sha256": "569f3c779e6a0b8210a764c9096e49e8a5199c107a64f7fcfb7a011257bf2a12", "supports": ["aliases"], "url": "https://api.github.com/repos/cooelf/Auto-GUI"}], "id": "github:cooelf/auto-gui", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/cooelf/Auto-GUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/cooelf/Auto-GUI"}], "name": "Auto-GUI / Auto-UI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/cooelf/Auto-GUI/HEAD/README.md", "target_id": "arxiv:2309.11436", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screenshot-based mobile agent implementation with training and inference code for chain-of-action prediction.", "topics": ["mobile-navigation", "planning"], "updated_at": null, "url": "https://github.com/cooelf/Auto-GUI"}
{"aliases": [], "authors": [], "details": {"availability": "Published source; official release metadata inspected", "release_tag": "v0.8.8"}, "environments": ["windows", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "bbec20828c8a0f5483be68f6a00c47b202e9431268d96c1195fbf0745aca1364", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/CursorTouch/Windows-MCP/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "ef43250ad8e57c3267f2a8e0aaeb655eb3445ade8d2dee8c243dc6e241a9d2d8", "supports": ["identity"], "url": "https://api.github.com/repos/CursorTouch/Windows-MCP"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "652b580c99ac2de9d56d124225ba302f0dd83430ccc9b5162e16f683c56b952f", "supports": ["release_status"], "url": "https://api.github.com/repos/CursorTouch/Windows-MCP/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:58Z", "role": "primary", "sha256": "db3de978027106a991da3f7f2d78b216cfffee9b08604872fa3216927acc7a2d", "supports": ["release_status"], "url": "https://api.github.com/repos/CursorTouch/Windows-MCP/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:59Z", "role": "primary", "sha256": "20ced59173768e20e91a43e3a80443aa1200b129c134f07f0cc45779747a1743", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/CursorTouch/Windows-MCP/HEAD/LICENSE.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:00:59Z", "role": "primary", "sha256": "20ced59173768e20e91a43e3a80443aa1200b129c134f07f0cc45779747a1743", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/CursorTouch/Windows-MCP/HEAD/LICENSE.md"}], "id": "github:cursortouch/windows-mcp", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T03:59:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T04:00:59Z", "url": "https://raw.githubusercontent.com/CursorTouch/Windows-MCP/HEAD/LICENSE.md"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/CursorTouch/Windows-MCP"}], "name": "Windows-MCP", "notes": "Public source and official release metadata are present. The README explicitly supports Windows 10 and 11 and includes a browser DOM inspection mode. Code is MIT-licensed. This is an execution tool, not a trained model; runtime behavior and latency claims were not independently tested.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A Windows MCP bridge exposing application, window, keyboard and mouse operations plus native UI-state inspection for agents.", "topics": ["tool-integration", "automation", "screen-understanding"], "updated_at": null, "url": "https://github.com/CursorTouch/Windows-MCP"}
{"aliases": [], "authors": [], "details": {"availability": "Public inference/scoring source; benchmark data access remains gated.", "evaluation_caution": "Missing screenshots are marked AUTO_PASS and credited as correct according to the README.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/LICENSE"], "reason": "Repository code license only; no data-license or gated-access permission is inferred."}}, "repository_revision": "0c2da7002baa5b0ee71f80dd6fbd90e33071e259"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:21Z", "role": "primary", "sha256": "6681ce1bd98c856fed4193e239e234a69eb2a74fa937e9c38db89f48db40f452", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:20Z", "role": "primary", "sha256": "f2b7c762383746a56e62273f756ca7ed3223726535763a41fda3ac5e3467aa82", "supports": ["identity"], "url": "https://api.github.com/repos/CVC2233/AndroTMem"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:59Z", "role": "primary", "sha256": "302c804dc65ebb3e8558651e0c7cd0a79c1caf29fc9fb11868d26fba36afdf48", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/CVC2233/AndroTMem/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:00Z", "role": "primary", "sha256": "6d572ccef64746baac78441c0b26488f525c70c2b7453f985bd4f19cbec00634", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:07Z", "role": "primary", "sha256": "e82623b7dc889eb7bff45c33f15a47c8c5caac3de0a3c56a7a9975ca9b2d6694", "supports": ["notes", "release_status"], "url": "https://huggingface.co/api/datasets/CVC2233/AndroTMem-Bench"}], "id": "github:cvc2233/androtmem", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T11:18:21Z", "licenses": {"code": {"basis": "Repository code license only; no data-license or gated-access permission is inferred.", "declared": "MIT", "retrieved_at": "2026-10-17T11:23:00Z", "url": "https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/CVC2233/AndroTMem"}], "name": "AndroTMem inference and evaluation code", "notes": "The source tree releases inference backends and evaluation scripts under evaluation/. The separately hosted benchmark is gated=auto and its raw card returned 401; no access request was made and it is not approved here as a directly accessible dataset. The README explicitly credits missing screenshots as AUTO_PASS/correct, so incomplete local image inventories can inflate scores. Its TCR is an offline anchor/dependency score and is not presented here as independently verified live task success.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/README.md", "target_id": "arxiv:2603.18429", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A public Android GUI evaluation framework with action-history, summary and anchored-memory prompt modes, plus step-action and dependency-aware task scoring.", "topics": ["memory", "long-horizon", "evaluation"], "updated_at": null, "url": "https://github.com/CVC2233/AndroTMem"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "33c55f7d1df7f812c9c9b9c4d97a4de7408b1b1916e58eed266b1f88389b50cd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/DigiRL-agent/digiq/HEAD/README.md"}], "id": "github:digirl-agent/digiq", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/DigiRL-agent/digiq"}], "name": "Digi-Q", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/DigiRL-agent/digiq/HEAD/README.md", "target_id": "arxiv:2502.15760", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Offline device-control training code that learns action values and uses them to select promising actions, with linked trajectories and checkpoints.", "topics": ["reinforcement-learning", "efficiency"], "updated_at": null, "url": "https://github.com/DigiRL-agent/digiq"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "464e26d88420cd2694abe4c014bfd088417622c3a980aadf19b2bf01988d5b65", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/DigiRL-agent/digirl/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/DigiRL-agent/digirl/HEAD/LICENSE"}], "id": "github:digirl-agent/digirl", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/DigiRL-agent/digirl/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/DigiRL-agent/digirl"}], "name": "DigiRL", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/DigiRL-agent/digirl/HEAD/README.md", "target_id": "arxiv:2406.11896", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Reinforcement learning for autonomous mobile device control.", "topics": ["reinforcement-learning"], "updated_at": null, "url": "https://github.com/DigiRL-agent/digirl"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "f501f8de6aac53758f60882a8c595a5ac27145f327c73b0cfa7921b45ccae75a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/DistRL-lab/distrl-open/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/DistRL-lab/distrl-open/HEAD/LICENSE"}], "id": "github:distrl-lab/distrl-open", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/DistRL-lab/distrl-open/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/DistRL-lab/distrl-open"}], "name": "DistRL", "notes": "The README identifies model weights and dataset release as pending; this entry covers the code framework.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/DistRL-lab/distrl-open/HEAD/README.md", "target_id": "arxiv:2410.14803", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Distributed online reinforcement-learning framework for Android agents, with environment setup and trajectory collection tools.", "topics": ["reinforcement-learning", "efficiency"], "updated_at": null, "url": "https://github.com/DistRL-lab/distrl-open"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:05Z", "role": "primary", "sha256": "3fb5aabfb07df620955f0b74347f9d5926a1cd3d506b1098eccba9477627269f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/dmis-lab/SwitchMem/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "cec8e4679c883f13983915e9f77d175a84a234d69386e78f9ea75157cefbdcb4", "supports": ["relations"], "url": "https://txxnrd.github.io/switchmem/"}], "id": "github:dmis-lab/switchmem", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T01:46:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/dmis-lab/SwitchMem"}], "name": "SwitchMem", "notes": "The implementation distinguishes step-level AITW/LearnGUI scores from terminal task success on MobileMiniWoB++. Benchmark outcomes and local smoke-test outputs are different claims; neither was independently reproduced here.", "published_at": null, "relations": [{"evidence_url": "https://txxnrd.github.io/switchmem/", "target_id": "arxiv:2610.17928", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Memory-routing implementation that chooses text or image demonstrations for a frozen GUI actor, with offline evaluation and Android transfer experiments.", "topics": ["memory", "mobile-navigation"], "updated_at": null, "url": "https://github.com/dmis-lab/SwitchMem"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md", "population": "dataset taxonomy", "qualifier": "exact", "unit": "GUI scenario categories", "value": 6}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md", "population": "dataset taxonomy", "qualifier": "exact", "unit": "question types", "value": 8}], "evidence_urls": ["https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md", "https://huggingface.co/datasets/ONE-Lab/GUI-World/raw/main/README.md", "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"], "formats": [], "modalities": ["videos", "frame images", "text"], "notes": "Taxonomy counts describe coverage, not the number of videos or questions. Cases include Android, iOS, desktop software, websites, multi-application use, and XR. Keyframe-selection settings affect evaluation; exact sample inventories are not reported here.", "splits": ["train", "test"], "supervision": ["GUI video question answering", "temporal interface understanding"]}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:52Z", "role": "primary", "sha256": "d02fa5da3384a3e58cf1d28aad47497de73fe3aad0ce57219585da04fc3187ef", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://github.com/Dongping-Chen/GUI-World"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:23Z", "role": "primary", "sha256": "905bf69ce80c866377841474cefe08f85ec4ef552cdd9ffeb2fb1531e07370b6", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:06:30Z", "role": "primary", "sha256": "d2dd5886bce90f6eaee880719c79a6ede6ec0eaa08da32d7610bde656b8dc9ed", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/ONE-Lab/GUI-World/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["details.dataset_profile"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "id": "github:dongping-chen/gui-world", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:40:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Dongping-Chen/GUI-World"}], "name": "GUI-World", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://github.com/Dongping-Chen/GUI-World", "target_id": "arxiv:2406.10819", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI video benchmark and training data covering six scenarios, from Android and iOS to websites and XR, with eight question types for image and video LLMs and the GUI-Vid model.", "topics": ["video", "screen-understanding"], "updated_at": null, "url": "https://github.com/Dongping-Chen/GUI-World"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "0d05433c730e74f13be753fa6156737f797a69d61d8793e47f5a56b3a63ee790", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/droidrun/mobile-harness/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "de215a2914d50cb87fcab9e19efc856b9e600a68eefec5c63e284dae34977133", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/droidrun/mobile-harness/HEAD/LICENSE"}], "id": "github:droidrun/mobile-harness", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/droidrun/mobile-harness/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/droidrun/mobile-harness"}], "name": "Mobile Harness", "notes": "The publisher describes this as a Markdown harness, not an agent runtime. Published benchmark results have not been independently reproduced.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Portable agent instructions and integration tooling for local or cloud phone control through mobilerun-core.", "topics": ["automation", "tool-integration"], "updated_at": null, "url": "https://github.com/droidrun/mobile-harness"}
{"aliases": ["Droidrun"], "authors": [], "details": {"availability": "Published source; official release metadata inspected", "release_tag": "v0.7.0"}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "80052560e06db353277a55f3f8f6a4b5231aa6dc1376d5182559b72adcb5c3a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/droidrun/mobilerun/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:56Z", "role": "primary", "sha256": "e48b9154a58b4b7021a4707ca185af32b9dd81c828d53ba36a86f195c1720344", "supports": ["identity"], "url": "https://api.github.com/repos/droidrun/mobilerun"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "7f40d42acd62b4405b5077a0f594a35aae73fb057589346afa5f6fb22d710a0c", "supports": ["release_status"], "url": "https://api.github.com/repos/droidrun/mobilerun/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:56Z", "role": "primary", "sha256": "f85c8b766edc107c455a86e363c56af3d1f7fce669b55377f4010982aa3a7e44", "supports": ["release_status"], "url": "https://api.github.com/repos/droidrun/mobilerun/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:57Z", "role": "primary", "sha256": "b3ebb19fda8286efa415d457529eafff41e1b1a27fe4e00d99f34e1559b6cdd9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/droidrun/mobilerun/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "e48b9154a58b4b7021a4707ca185af32b9dd81c828d53ba36a86f195c1720344", "supports": ["identity", "aliases", "notes"], "url": "https://api.github.com/repos/droidrun/droidrun"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:00:57Z", "role": "primary", "sha256": "b3ebb19fda8286efa415d457529eafff41e1b1a27fe4e00d99f34e1559b6cdd9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/droidrun/mobilerun/HEAD/LICENSE"}], "id": "github:droidrun/mobilerun", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T03:59:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T04:00:57Z", "url": "https://raw.githubusercontent.com/droidrun/mobilerun/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/droidrun/mobilerun"}], "name": "Mobilerun (formerly Droidrun)", "notes": "The former droidrun/droidrun repository resolves to droidrun/mobilerun through the official API. Source and release v0.7.0 are public. Android and iOS setup paths differ, and the companion Portal app is a separate device bridge. MIT applies to framework code; model-provider and cloud-service terms are separate.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A framework that combines screenshots and accessibility state with agent tools for mobile-device tasks through command-line and Python interfaces.", "topics": ["mobile-navigation", "tool-integration", "planning"], "updated_at": null, "url": "https://github.com/droidrun/mobilerun"}
{"aliases": ["Droidrun Portal"], "authors": [], "details": {"availability": "Published source; official release metadata inspected", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/droidrun/mobilerun-portal/HEAD/LICENSE"], "reason": "GitHub classifies this license as NOASSERTION, but the actual program-specific notice is explicit. Do not reduce the permission to a generic unknown or silently transfer the framework MIT declaration."}}, "release_tag": "v0.7.25"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "fd2257dd735dbb755f2de36b44e425e5862004a04130e53e92b330f1419479a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/droidrun/mobilerun-portal/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:56Z", "role": "primary", "sha256": "be519be28e277bd979717996e9587879c6a5d9235b95944305252b18595da413", "supports": ["identity"], "url": "https://api.github.com/repos/droidrun/mobilerun-portal"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "65e81cfc9f227350e33da9e09030a7098e82a6021aa9b04cf0485ed2acee95a0", "supports": ["release_status"], "url": "https://api.github.com/repos/droidrun/mobilerun-portal/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:01Z", "role": "primary", "sha256": "31b8d75541bbfb7dba2f80adc83aa099ef83ca057b467eb2cb98b586f6d9d045", "supports": ["release_status"], "url": "https://api.github.com/repos/droidrun/mobilerun-portal/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:01Z", "role": "primary", "sha256": "945331c1cfa589694eb6ccc51574a72ee215a01b10bd5556f1b8be8d11bd71ab", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/droidrun/mobilerun-portal/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "be519be28e277bd979717996e9587879c6a5d9235b95944305252b18595da413", "supports": ["identity", "aliases", "notes"], "url": "https://api.github.com/repos/droidrun/droidrun-portal"}], "id": "github:droidrun/mobilerun-portal", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": {"basis": "The program-specific copyright notice explicitly permits GNU AGPL version 3 or any later version.", "declared": "AGPL-3.0-or-later", "retrieved_at": "2026-10-17T04:01:01Z", "url": "https://raw.githubusercontent.com/droidrun/mobilerun-portal/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/droidrun/mobilerun-portal"}], "name": "Mobilerun Portal (formerly Droidrun Portal)", "notes": "The official API resolves the former Droidrun Portal path to mobilerun-portal. Release metadata lists a v0.7.25 Android APK. Although GitHub labels its license NOASSERTION, the inspected LICENSE explicitly grants AGPL version 3 or later for the program. Accessibility-dependent UI functions require the service; no permissions were granted or device operated.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "An Android accessibility service that exposes UI-element inspection, device-control interfaces and screen feedback to mobile automation agents.", "topics": ["mobile-navigation", "screen-understanding", "tool-integration"], "updated_at": null, "url": "https://github.com/droidrun/mobilerun-portal"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "edb4293cb510888cdc2d5c9d1135a0500272aae96366710744d5b74b5289c462", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/e2b-dev/desktop/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/e2b-dev/desktop/HEAD/LICENSE"}], "id": "github:e2b-dev/desktop", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/e2b-dev/desktop/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/e2b-dev/desktop"}], "name": "E2B Desktop", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/e2b-dev/desktop/HEAD/README.md", "target_id": "organization:e2b", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Open-source, isolated Linux Xfce desktop sandboxes for computer-use agents, driven from Python or JavaScript SDKs with screen streaming, mouse, keyboard and window control, screenshots and shell.", "topics": ["sandbox"], "updated_at": null, "url": "https://github.com/e2b-dev/desktop"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "fc3099072ca60acf2294a63dd2c3e343e90c901a7c7581f723cfdccc5a1edc2b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/EmergenceAI/Agent-E/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "555ac6a8bf7ef4a3c82f4ee258be9d751042f8d2a6c901e5eac7c43596563c75", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/EmergenceAI/Agent-E/HEAD/LICENSE.txt"}], "id": "github:emergenceai/agent-e", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/EmergenceAI/Agent-E/HEAD/LICENSE.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/EmergenceAI/Agent-E"}], "name": "Agent E", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/EmergenceAI/Agent-E/HEAD/README.md", "target_id": "arxiv:2407.13032", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Browser automation agent on the AG2 framework that pairs a planner agent with a browser navigation agent to fill forms, search shops and work through sites, also callable over a local HTTP API.", "topics": ["web-navigation"], "updated_at": null, "url": "https://github.com/EmergenceAI/Agent-E"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md", "https://api.github.com/repos/Euphoria16/UI-Genie/git/trees/main?recursive=1"], "reason": "The README declares MIT and links a root LICENSE, but the complete retained tree has no root LICENSE; bundled ms-swift has its own license. Preserve this unresolved scope instead of treating a component license as the whole project license."}}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "e08569bd9d8a6ff44537d4fe69243ddfeb8a41ae5c9f87ef84af50a67c64c726", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "a53e4cd3b4391c03c9334745152f8c1f66aff7590051f9bdb7e78b81c6c99a24", "supports": ["identity"], "url": "https://api.github.com/repos/Euphoria16/UI-Genie"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:55Z", "role": "primary", "sha256": "c27c6d88d043c964776c04a74517d48644418dca2f5ae9c8ba778d082c12a6ae", "supports": ["release_status", "licenses.code"], "url": "https://api.github.com/repos/Euphoria16/UI-Genie/git/trees/main?recursive=1"}], "id": "github:euphoria16/ui-genie", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Euphoria16/UI-Genie"}], "name": "UI-Genie", "notes": "The tree contains ms-swift-based training and AndroidControl evaluation scripts. The README says AndroidLab/Android Arena scripts and the 72B checkpoint are still forthcoming; these are not claimed released. It declares MIT but links a root LICENSE absent from the inspected tree, so the exact code-license scope remains unresolved. Separate Hub weights and data have their own declarations.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md", "target_id": "arxiv:2505.21496", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training and evaluation code for a mobile-agent framework that iteratively generates trajectories using agent and reward-model feedback.", "topics": ["data-generation", "reward-models", "imitation-learning"], "updated_at": null, "url": "https://github.com/Euphoria16/UI-Genie"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:34Z", "role": "primary", "sha256": "9c938f8f3f3ee04f459362a8b7d399588fd030ce476cda111e5f23fbf1e31f10", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/facebookresearch/ai-agent-privacy/HEAD/README.md"}], "id": "github:facebookresearch/ai-agent-privacy", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/facebookresearch/ai-agent-privacy"}], "name": "AgentDAM", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/facebookresearch/ai-agent-privacy/HEAD/README.md", "target_id": "arxiv:2503.09780", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Web-agent benchmark for measuring unnecessary use of sensitive personal information during end-to-end task execution.", "topics": ["safety-security", "evaluation"], "updated_at": null, "url": "https://github.com/facebookresearch/ai-agent-privacy"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["click", "type", "scroll", "configured BrowserGym/AgentLab actions"], "counts": [], "evaluation": "Tasks define check_if_task_is_complete(initial_state,current_state); app-specific reward logic evaluates underlying state, with custom checks for code-editor file contents. Observation/history flags are configurable. No universal single fixed task total or published-run reproduction is implied.", "evidence_urls": ["https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/README.md", "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/docs/tasks.md", "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/docs/agents.md", "https://api.github.com/repos/facebookresearch/OpenApps/git/trees/main?recursive=1"], "focus": ["UI-agent reliability under environment variation", "state-based browser task evaluation"], "notes": "Preserve the code/tag, app state, theme/layout/content choices, goal variant and action/observation configuration. Current main refactors the paper-era appearance surface, and the v1.0-paper tag includes post-publication fixes that may move rewards. The environments run as browser apps; phone/tablet viewports are not native Android/iOS environments.", "observations": ["screenshots", "accessibility trees", "set-of-marks and coordinates (configurable)"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "6e6782794433c60468d65ac75bc3f03c56007f8c", "repository_url": "https://github.com/facebookresearch/OpenApps"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:18Z", "role": "primary", "sha256": "1df0f66584b37e7dba90811a7732d3b0c13c96cd7da4aec3970a18be36a6cec1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:19Z", "role": "primary", "sha256": "2c57b627cb344e360480c10e74dcdd1d1ed321d9a7a4e3891bc970355b15f4cd", "supports": ["identity"], "url": "https://api.github.com/repos/facebookresearch/OpenApps"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:18Z", "role": "primary", "sha256": "34aa76fbd6cca6ac2f3949792144662b90d85d7a18988afabcb570167ea5c1d6", "supports": ["identity", "release_status", "details.repository_tree_sha", "details.benchmark_protocol"], "url": "https://api.github.com/repos/facebookresearch/OpenApps/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:40Z", "role": "primary", "sha256": "af030b93a7604c063b08f003b0a0124a6e95e560649d0306bfe40dd87fed563d", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/docs/tasks.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:40Z", "role": "primary", "sha256": "aede4dfd965bd3484336f40ddb1bd6db48aed00ab45e27c048b76cc4601d71b3", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/docs/agents.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:20Z", "role": "primary", "sha256": "336255dc30193e8e15d689d9481bb05673d89055718f3a96923a7ffb99adbbaf", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/LICENSE"}], "id": "github:facebookresearch/openapps", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T04:35:18Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "CC-BY-NC-4.0", "retrieved_at": "2026-10-18T04:35:20Z", "url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/facebookresearch/OpenApps"}], "name": "OpenApps", "notes": "Public code includes app/task/configuration and agent-launching implementations. README distinguishes moving main from v1.0-paper and explicitly says even that tag is not a byte-exact snapshot of the published runs. CC-BY-NC-4.0 is the repository’s declared code license; third-party WebShop, map and icon content retain separate terms. No commercial-use permission or fixed downloadable task total is inferred.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/README.md", "target_id": "arxiv:2511.20766", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Configurable browser apps and task rewards for testing computer-use agents under changes in content, layout and appearance.", "topics": ["evaluation", "robustness", "cross-app", "sandbox"], "updated_at": null, "url": "https://github.com/facebookresearch/OpenApps"}
{"aliases": ["STEVE", "STEVE-R1"], "authors": [], "details": {}, "environments": ["desktop", "windows", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:14:29Z", "role": "primary", "sha256": "52aa383fdf5d91951f5d9a73f4ebed4454a82b35edae296918116e9f0895dd83", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/FanbinLu/STEVE/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["relations", "environments"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "github:fanbinlu/steve-r1", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T13:14:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/FanbinLu/STEVE-R1"}], "name": "STEVE / STEVE-R1 research code", "notes": "The repository has moved to STEVE-R1. Its README reports an SFT release, with RL tuning in progress, and describes evaluations using multiple attempts per task. These results are not independently reproduced here.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2503.12532", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Research code and linked models for step-verified computer-use training; the current repository describes STEVE-R1 and its trajectory releases.", "topics": ["data-generation", "automation", "evaluation"], "updated_at": null, "url": "https://github.com/FanbinLu/STEVE-R1"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["configurable coordinate actions", "element-reference actions", "keyboard input"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md", "population": "README collection size", "qualifier": "more-than", "unit": "web environments", "value": 100}], "evaluation": "", "evidence_urls": ["https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md", "https://miniwob.farama.org/content/observation_space/", "https://miniwob.farama.org/content/action_space/"], "focus": ["small web interaction tasks"], "notes": "The README gives a lower bound, not an exact active-environment inventory. Action presets expose different capabilities, so comparisons must identify the preset. Environment-specific reward details are not normalized here. The repository is in maintenance mode.", "observations": ["task instructions", "screenshots", "visible DOM-element attributes"], "setting": "interactive"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "90499776fa1b23e08ac85190ad3be709caa9172ff4f2a0a106b0fa350568e86f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "90499776fa1b23e08ac85190ad3be709caa9172ff4f2a0a106b0fa350568e86f", "supports": ["details.benchmark_protocol", "relations", "notes"], "url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:23Z", "role": "primary", "sha256": "511039e0e3c26b117e7e0a001746c26ec6d94a8ab8c467f0be0bc1a0b01e870a", "supports": ["details.benchmark_protocol"], "url": "https://miniwob.farama.org/content/observation_space/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:23Z", "role": "primary", "sha256": "ae7bf6a3b98b198b985d7ca7b41cb2bdab3cf48e6445467ad9a8d4a2c8df240f", "supports": ["details.benchmark_protocol"], "url": "https://miniwob.farama.org/content/action_space/"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "fd6529fb8f648d4130fd969e3d7fc740c9992a0bd94a1998f52555a873311511", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/LICENSE"}], "id": "github:farama-foundation/miniwob-plusplus", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Farama-Foundation/miniwob-plusplus"}], "name": "MiniWoB++", "notes": "The maintainer README identifies MiniWoB++ as the extension introduced in the 2018 workflow-guided exploration paper. It separately links World of Bits as the source for the original MiniWoB predecessor. The current Farama interface, environment inventory, and maintenance status belong to the retained repository snapshot.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md", "target_id": "arxiv:1802.08802", "type": "introduced_by"}], "review_status": "curated", "scope": "core", "summary": "Over 100 web-interaction environments extending OpenAI's MiniWoB, controlled from Python through a Gymnasium API that drives the browser with Selenium.", "topics": ["evaluation", "reinforcement-learning"], "updated_at": null, "url": "https://github.com/Farama-Foundation/miniwob-plusplus"}
{"aliases": ["github:mobilegptsys/mobilegpt", "mobilegptsys/MobileGPT", "https://github.com/mobilegptsys/MobileGPT"], "authors": [], "details": {"availability": "Public Android client, Python server and manual offline-explorer source.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/fclab-skku/MobileGPT/HEAD/LICENSE"], "reason": "Repository source license; linked benchmark data and external models retain separate terms."}}, "repository_revision": "7945048e6c05849aff1633744cbdf653c58bece2"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "8683cff296d21c1a7a38bc372efaf004b8af644030fac77a4dc9d2b2e21d894b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/fclab-skku/MobileGPT/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "3d3ceb084d374b706842cff1f4ec986fde50c13cba141dc4619d5afe8b631d55", "supports": ["identity"], "url": "https://api.github.com/repos/mobilegptsys/MobileGPT"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "2a8f30b3422a033b345c2054d55adff0836b7dac163e600583ba0db1b4423320", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/fclab-skku/MobileGPT/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/fclab-skku/MobileGPT/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:04Z", "role": "primary", "sha256": "3d3ceb084d374b706842cff1f4ec986fde50c13cba141dc4619d5afe8b631d55", "supports": ["aliases"], "url": "https://api.github.com/repos/mobilegptsys/MobileGPT"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:06Z", "role": "primary", "sha256": "3d3ceb084d374b706842cff1f4ec986fde50c13cba141dc4619d5afe8b631d55", "supports": ["aliases"], "url": "https://api.github.com/repos/fclab-skku/MobileGPT"}], "id": "github:fclab-skku/mobilegpt", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T12:32:07Z", "licenses": {"code": {"basis": "Repository source license; linked benchmark data and external models retain separate terms.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/fclab-skku/MobileGPT/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/fclab-skku/MobileGPT"}], "name": "MobileGPT", "notes": "The old mobilegptsys repository resolves to fclab-skku/MobileGPT. Public Android client and Python server sources require Android SDK 33+, Accessibility Service setup and external service credentials. The README explicitly omits the random explorer and human-in-the-loop memory repair from the open-source version. Its benchmark description differs from the retained arXiv snapshot, so no dataset counts or reproduction claim are merged into this code record. External benchmark data was not approved as a separate release.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50", "target_id": "arxiv:2312.03003", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An Android client and Python-server agent that learns reusable app subtasks through exploration, selection, derivation and memory recall.", "topics": ["mobile-navigation", "memory", "planning", "workflow"], "updated_at": null, "url": "https://github.com/fclab-skku/MobileGPT"}
{"aliases": [], "authors": [], "details": {"availability": "Public evaluation harness and ungated web-application archive manifest.", "benchmark_protocol": {"actions": ["browser interaction", "test checklist generation", "defect reports"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/README.md", "population": "publisher benchmark overview", "qualifier": "exact", "unit": "application categories", "value": 7}, {"basis": "computed", "evidence_url": "https://huggingface.co/api/datasets/friedrichor/WebTestBench", "population": "Hub file manifest", "qualifier": "exact", "unit": "web-application archives", "value": 100}], "evaluation": "The supplied WebTester pipeline separates checklist generation from interactive defect detection and provides a scoring workflow.", "evidence_urls": ["https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/README.md", "https://huggingface.co/api/datasets/friedrichor/WebTestBench", "https://huggingface.co/datasets/friedrichor/WebTestBench/raw/main/README.md"], "focus": ["web functionality testing", "latent permission and business constraints"], "notes": "The 100 archive files are not counted as 100 independent test tasks or defects. Backend/tool versions and deployment settings affect reproducibility.", "observations": ["web application interface", "development instruction"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/LICENSE"], "reason": "Repository source LICENSE; external tools and services retain their own licenses."}, "data": {"evidence_urls": ["https://huggingface.co/datasets/friedrichor/WebTestBench/raw/main/README.md"], "reason": "Publisher Hub-card license declaration for the released benchmark dataset."}}, "repository_revision": "04873065bdb0938ed53d809559d4ab473c7ccd31"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:18Z", "role": "primary", "sha256": "b41472026cd7d4f81a01d58bf39193dc1fabc7875c61e370090a20cc77500cdd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:18Z", "role": "primary", "sha256": "7203ecc2aae5f1697eeeb804ca8f21d47b9f22a8b8fa43a4064c4e1c49f5e72c", "supports": ["identity"], "url": "https://api.github.com/repos/friedrichor/WebTestBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:55Z", "role": "primary", "sha256": "bd1f09149b00d59a0078d414216bd2fb78e09ce7212d15bbda84ddec7052c8af", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/friedrichor/WebTestBench/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:08Z", "role": "primary", "sha256": "db1e97f40107038ae03835dc20f17b13b37e74b5c8587f8c39b441271618f017", "supports": ["identity", "release_status", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/friedrichor/WebTestBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:08Z", "role": "primary", "sha256": "6f2f94a25ed66a14779f6b4e6686044d0510eb32a0187645cc473978d93f7c5b", "supports": ["licenses.data", "details.benchmark_protocol"], "url": "https://huggingface.co/datasets/friedrichor/WebTestBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:55Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/LICENSE"}], "id": "github:friedrichor/webtestbench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T11:18:18Z", "licenses": {"code": {"basis": "Repository source LICENSE; external tools and services retain their own licenses.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T11:22:55Z", "url": "https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/LICENSE"}, "data": {"basis": "Publisher Hub-card license declaration for the released benchmark dataset.", "declared": "apache-2.0", "retrieved_at": "2026-10-17T11:23:08Z", "url": "https://huggingface.co/datasets/friedrichor/WebTestBench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/friedrichor/WebTestBench"}], "name": "WebTestBench", "notes": "Public evaluation/deployment scripts accompany an ungated Hub release with JSON/JSONL metadata and 100 application ZIP files. The ZIP count is an artifact inventory, not an asserted test-case or defect count. The documented baseline uses Claude Agent SDK and a pinned Playwright MCP version with external model credentials; these were not installed or run. Code and data Apache-2.0 declarations are retained separately.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/README.md", "target_id": "arxiv:2603.25226", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An end-to-end web-testing benchmark where agents derive test checklists from development instructions, interact with deployed applications and report functional or constraint-related defects.", "topics": ["evaluation", "web-navigation", "tool-integration"], "updated_at": null, "url": "https://github.com/friedrichor/WebTestBench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "178774927f003b5221d695fa78ded6fcc543aa8653c7b2dd8016418e6911afb7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "aee22091f60f10968afb7dc33dd8b728ea929e287c260ed82adffdec4866fcb5", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent/HEAD/LICENSE"}], "id": "github:gair-nlp/pc-agent", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/GAIR-NLP/PC-Agent"}], "name": "PC-Agent", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent/HEAD/README.md", "target_id": "arxiv:2412.17589", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Human-cognition-transfer framework for PC agents: PC Tracker logs human computer use, cognition completion turns logs into cognitive trajectories, and a planner works with a grounding agent.", "topics": ["planning", "cross-app"], "updated_at": null, "url": "https://github.com/GAIR-NLP/PC-Agent"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "9b1b00b907b99a4a8dbfc9c01ea49010878dfbd6", "repository_url": "https://github.com/GAIR-NLP/PC-Agent-E"}, "environments": ["desktop", "windows", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:13Z", "role": "primary", "sha256": "3ffb8c843c6acb8c2d5bbba11ba49d6b36cd566553a89974b6252a2bfd289d5d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:01Z", "role": "primary", "sha256": "30761141ea62c5f20183f359bbc3fc8ccc5b458886d38c5426bae4802a23ee0b", "supports": ["identity"], "url": "https://api.github.com/repos/GAIR-NLP/PC-Agent-E"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:23Z", "role": "primary", "sha256": "2fb2aa4ee6f02ccadb631dab4972f96c3fdf545cd92c5854fc191e81cf4ae88a", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/GAIR-NLP/PC-Agent-E/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "39e37b84c067596361c491e2e9e22af8805386a24b8352b87e9aca3bddf4f92c", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/LICENSE"}], "id": "github:gair-nlp/pc-agent-e", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:13Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/GAIR-NLP/PC-Agent-E"}], "name": "PC-Agent-E training and deployment", "notes": "README demonstrates Windows and Linux operation and describes thought completion and trajectory augmentation. The released source tree contains postprocess, train and deploy code. The checkpoint, data archive and WindowsAgentArena-V2 are separate resources; external model APIs and a training framework are required.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md", "target_id": "arxiv:2505.13909", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Tools for refining and augmenting human computer-use trajectories, preparing supervised training data, and deploying a native PC-Agent-E scaffold.", "topics": ["data-generation", "imitation-learning", "automation"], "updated_at": null, "url": "https://github.com/GAIR-NLP/PC-Agent-E"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["PyAutoGUI"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md", "population": "full WindowsAgentArena-V2 suite reported by README", "qualifier": "exact", "unit": "tasks", "value": 141}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md", "population": "full suite coverage reported by README", "qualifier": "exact", "unit": "Windows applications", "value": 11}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_custom.json", "population": "shipped test_custom.json selection", "qualifier": "exact", "unit": "task IDs", "value": 20}], "evaluation": "Runner exposes four observation choices and uses PyAutoGUI. count.py sums the last score in each results.json and divides by the number of traj.jsonl files, multiplying by 100; it separately reports result/trajectory/error-file coverage. This denominator is not automatically the full benchmark size.", "evidence_urls": ["https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md", "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/run.py", "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_custom.json", "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/count.py", "https://api.github.com/repos/GAIR-NLP/WindowsAgentArena-V2/git/trees/main?recursive=1"], "focus": ["Windows desktop task execution", "revised task evaluators"], "notes": "README sample config uses max_steps=30. The shipped selected-task index has 20 IDs, not 141; preserve the selected task list, trial ID, score coverage and VM snapshot when comparing runs. Docker/Windows-11 setup is required. No VM or benchmark was executed.", "observations": ["screenshot", "accessibility tree", "screenshot plus accessibility tree", "set-of-marks"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "2927fe55005d1be75d5a9188c0045f73ba28d192", "repository_url": "https://github.com/GAIR-NLP/WindowsAgentArena-V2"}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:17Z", "role": "primary", "sha256": "1dd744ce565ec00b20ea5637eb7bc6112e97fb9044bd9a471f7cc4b7227ca344", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:20Z", "role": "primary", "sha256": "0b0af02d3c2c4916dd786b868b99d60b4681927d0273a23661eb2b6db7cf9404", "supports": ["identity"], "url": "https://api.github.com/repos/GAIR-NLP/WindowsAgentArena-V2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:16Z", "role": "primary", "sha256": "638ef9ee593a3a560c1d0b3fb2bc4a9d31a3e904c0111971924a6d66e1763aa6", "supports": ["identity", "release_status", "details.repository_tree_sha", "details.benchmark_protocol"], "url": "https://api.github.com/repos/GAIR-NLP/WindowsAgentArena-V2/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:39Z", "role": "primary", "sha256": "494de9e74629ac2c55358629c0d9ba13703f36317057a3089388fbb6e80ca28f", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/run.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:39Z", "role": "primary", "sha256": "e45ba7fdc7555f46ad1b4cfaab79d00754909d053fc49d5bab651b0fb593439b", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_custom.json"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:39Z", "role": "primary", "sha256": "f12147ac78c39c72eb75f9c37b293ba7fff3209fc5758fd51c69abbd96d80a97", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/count.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:20Z", "role": "primary", "sha256": "2a2d1e46940b570d60f5e2198ed27fcec6316fc8489e9660da8777dbd75aaeec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/LICENSE"}], "id": "github:gair-nlp/windowsagentarena-v2", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T04:35:17Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:20Z", "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/GAIR-NLP/WindowsAgentArena-V2"}], "name": "WindowsAgentArena-V2", "notes": "Repository contains executable evaluation code and task definitions. README links a separate Windows 11 VM snapshot; that image and its software licenses/access were not validated by downloading it. MIT is recorded for repository code only. Default task selection and the reported full suite have different scopes.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md", "target_id": "arxiv:2505.13909", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Windows benchmark revision with task/evaluator fixes and VM-initialization support, documented alongside PC-Agent-E.", "topics": ["evaluation", "cross-app", "robustness"], "updated_at": null, "url": "https://github.com/GAIR-NLP/WindowsAgentArena-V2"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "2e42f01578710d91145412f7d1ecb93979db96c698f677915284da1a1d592588", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/GaryYang12345/GUI-HARVEST/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/GaryYang12345/GUI-HARVEST/HEAD/LICENSE"}], "id": "github:garyyang12345/gui-harvest", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/GaryYang12345/GUI-HARVEST/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/GaryYang12345/GUI-HARVEST"}], "name": "GUI-HARVEST", "notes": "Requires a separately installed benchmark and seed agent harness; benchmark tasks and model weights remain fixed during its optimization procedure.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2610.00948", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Optimizes GUI-agent execution code from repeated rollout evidence and checks proposed changes against fixed validation gates.", "topics": ["automation", "observability", "robustness"], "updated_at": null, "url": "https://github.com/GaryYang12345/GUI-HARVEST"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/LICENSE"], "reason": "MIT declaration in the author repository LICENSE; separate data, weights and document scopes are not inferred."}}, "repository_id": 1079441024, "repository_revision": "2d69eeaf0ac44210c76dcab2d08b4d4610be236c"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:53.536297+00:00", "role": "primary", "sha256": "cbb08803922b4b7736979ca0a5fe552b111fcd428fa9317401bc090b14bb4e54", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:30.768857+00:00", "role": "primary", "sha256": "ff80afd35c79d8dd27dd7b9284ce082fe238c13a113aa19400ca0dc2d18ff194", "supports": ["identity", "details.repository_id"], "url": "https://api.github.com/repos/ghadinehme/VideoCAD"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:52.564274+00:00", "role": "primary", "sha256": "9ded7c6ef25b88879e9d46939ca7ff3bef97362e76dbdb808e57c67008541557", "supports": ["release_status", "details.repository_revision", "notes"], "url": "https://api.github.com/repos/ghadinehme/VideoCAD/git/trees/2d69eeaf0ac44210c76dcab2d08b4d4610be236c?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:34.311949+00:00", "role": "primary", "sha256": "a303d95a811331cd6ee23b0340f1b9ef2dd996835bfbc66656a871ff94eead50", "supports": ["environments", "summary", "relations", "notes"], "url": "https://arxiv.org/html/2505.24838v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:03:24.765415+00:00", "role": "primary", "sha256": "8e02753a1016dc67c97b395934313027b5f91b74209c7a24438de70fdc19cad1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/LICENSE"}], "id": "github:ghadinehme/videocad", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T12:02:53.536297+00:00", "licenses": {"code": {"basis": "The author repository at the inspected commit contains the standard MIT license. This declaration does not establish terms for externally hosted data or weights.", "declared": "MIT", "retrieved_at": "2026-10-18T12:03:24.765415+00:00", "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ghadinehme/VideoCAD"}], "name": "VideoCAD training and evaluation code", "notes": "The paper links this author repository. Training, preprocessing, model and evaluation source files are released; runtime behavior has not been independently reproduced. Onshape is a browser-based CAD environment; supported operating systems are not inferred. The full dataset and a checkpoint are linked separately, and their payloads were not downloaded for this review. MIT is recorded only for this repository code; external data and weight licenses remain unknown.", "published_at": null, "relations": [{"evidence_url": "https://arxiv.org/html/2505.24838v2", "target_id": "arxiv:2505.24838", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "VideoCADFormer training, action-log preprocessing and evaluation code for learning CAD interface action sequences from screen recordings of browser-based Onshape.", "topics": ["grounding", "imitation-learning", "evaluation"], "updated_at": null, "url": "https://github.com/ghadinehme/VideoCAD"}
{"aliases": ["github:deepmind/android_env", "deepmind/android_env", "https://github.com/deepmind/android_env"], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "b8dd2ea004647008265f4fb5c5a122569992311b6758f413080984266e53c25a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/google-deepmind/android_env/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "58d1e17ffe5109a7ae296caafcadfdbe6a7d176f0bc4ab01e12a689b0499d8bd", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/google-deepmind/android_env/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T15:37:19Z", "role": "primary", "sha256": "de82c4101e81a6980789904be59c3733ab13f2c1f0e769e5fcb1ab35f847cb3e", "supports": ["aliases"], "url": "https://api.github.com/repos/deepmind/android_env"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T15:37:20Z", "role": "primary", "sha256": "de82c4101e81a6980789904be59c3733ab13f2c1f0e769e5fcb1ab35f847cb3e", "supports": ["aliases"], "url": "https://api.github.com/repos/google-deepmind/android_env"}], "id": "github:google-deepmind/android_env", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/google-deepmind/android_env/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-deepmind/android_env"}], "name": "AndroidEnv", "notes": "Formerly published as deepmind/android_env. That address now redirects to this repository, and both paths serve the same reviewed README, so the project is counted once.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-deepmind/android_env/HEAD/README.md", "target_id": "arxiv:2105.13231", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Android device-control platform with screenshots, accessibility telemetry, input actions, and an optional reinforcement-learning task interface.", "topics": ["sandbox", "reinforcement-learning", "automation"], "updated_at": null, "url": "https://github.com/google-deepmind/android_env"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md", "population": "screen2words release", "qualifier": "exact", "unit": "summaries", "value": 112085}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md", "population": "RICO screens referenced by screen2words", "qualifier": "exact", "unit": "screens", "value": 22417}], "evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md"], "formats": ["CSV"], "modalities": ["text summaries", "RICO screen references"], "notes": "Multiple summaries describe one screen. The repository provides annotations and split files; use the referenced RICO assets for screenshots.", "splits": ["train", "dev", "test"], "supervision": ["screen summarization"]}, "modalities": ["text", "screenshot-reference"], "sample_count": 112085, "sample_unit": "screen summaries", "screen_count": 22417}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "8be66dc5a3907e0e84993a66679416d928ed1a32aae70db8bb0e0f771c7a8715", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.screen_count", "details.modalities", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md"}], "id": "github:google-research-datasets/screen2words", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research-datasets/screen2words"}], "name": "Screen2Words", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Mobile screen descriptions for language-based screen understanding.", "topics": ["screen-understanding"], "updated_at": null, "url": "https://github.com/google-research-datasets/screen2words"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md", "population": "original ScreenQA train split", "qualifier": "exact", "unit": "question-answer pairs", "value": 68951}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md", "population": "original ScreenQA validation split", "qualifier": "exact", "unit": "question-answer pairs", "value": 8614}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md", "population": "original ScreenQA test split", "qualifier": "exact", "unit": "question-answer pairs", "value": 8419}], "evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md"], "formats": ["JSON"], "modalities": ["questions", "answers", "screen references", "bounding boxes"], "notes": "All questions for a given screenshot share one split. ScreenQA Short reuses the same questions/screens; ComplexQA is a separate extension. Do not add these variants as independent original examples.", "splits": ["train", "validation", "test"], "supervision": ["screen question answering", "answer grounding"]}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "7d1bfe7d6eddd5d3fc4c33d4814de3cabf7a6b6fd3d3f85ba4c2b33036bf8867", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md"}], "id": "github:google-research-datasets/screen_qa", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research-datasets/screen_qa"}], "name": "ScreenQA", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md", "target_id": "arxiv:2209.08199", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "About 86,000 human-written question-answer pairs over 35,000 Rico mobile screenshots, plus ScreenQA Short and ComplexQA variants, for benchmarking screen content understanding.", "topics": ["screen-understanding"], "updated_at": null, "url": "https://github.com/google-research-datasets/screen_qa"}
{"aliases": [], "authors": [], "details": {"availability": "Public archived TFRecord annotation files for both downstream tasks.", "dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"], "formats": ["TFRecord"], "modalities": ["text", "bounding boxes", "UI hierarchy metadata", "Rico image identifiers"], "notes": "Original UI images are retrieved separately by Rico IDs. No release-specific sample count is established.", "splits": ["AppSim train/dev/test", "RefExp train/dev/test"], "supervision": ["similar UI element retrieval", "referring-expression component retrieval"]}, "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"], "reason": "The README explicitly says the datasets are licensed under CC BY 4.0; original Rico materials keep their source terms."}}, "repository_revision": "eceafbde7567959899bfb7f8d9e6fd71a52aaa12"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "45d1ca0f5c727cf58b2671dc5579f0628f4d78e07b977a5d8b018295c6fe1f58", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations", "details.dataset_profile", "licenses.data"], "url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "24a7838d425bdae363a7d5872702996b22068c1f0e47be274fdfbc4b0e20144f", "supports": ["identity"], "url": "https://api.github.com/repos/google-research-datasets/uibert"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "9a6fc19c6b407309943fe645f4addb3d032f153856976c4f8d7a8cfb6d2ffe6b", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/google-research-datasets/uibert/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "f752ebfb2598b7f848d343f428502b28e0f591f27849ab707f14bee398d4eb8d", "supports": ["environments", "notes"], "url": "https://interactionmining.org/rico"}], "id": "github:google-research-datasets/uibert", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": {"basis": "The README explicitly says the datasets are licensed under CC BY 4.0; original Rico materials keep their source terms.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-17T12:30:54Z", "url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research-datasets/uibert"}], "name": "UIBert AppSim and RefExp annotations", "notes": "The archived repository retains TFRecord train/dev/test files for AppSim and RefExp. These are annotation and UI-structure records with Rico image identifiers, not a republished complete screenshot collection or a UIBert checkpoint. The README’s 72k figure describes the upstream Rico source, not the number of released AppSim/RefExp examples; no sample total is inferred. The publisher explicitly assigns CC-BY-4.0 to these datasets.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md", "target_id": "url:0b134b082a7bd507fc6a", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md", "target_id": "arxiv:2107.13731", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Two Rico-derived UI component retrieval datasets: AppSim pairs semantically similar elements across screens, while RefExp links referring expressions to target elements.", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": null, "url": "https://github.com/google-research-datasets/uibert"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md", "population": "current three-annotator release", "qualifier": "exact", "unit": "mobile UI screens", "value": 1000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md", "population": "current three-annotator release", "qualifier": "exact", "unit": "design critiques", "value": 11344}], "evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md"], "formats": ["CSV"], "modalities": ["text", "bounding boxes", "quality ratings"], "notes": "Publisher-reported current release, separate from the smaller original-paper population. Screenshot pixels are obtained through referenced RICO IDs.", "splits": [], "supervision": ["design critique", "region localization", "quality rating"]}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:45Z", "role": "primary", "sha256": "345ff7003120f146016741099dd57b9b1a9ffc54ce1fc90f4d16d61e31d379e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:45Z", "role": "primary", "sha256": "5579bbbc622a2f9fa1f77b509c1ece17af023b8c14b53aa48dcc41288a5852f2", "supports": ["identity"], "url": "https://api.github.com/repos/google-research-datasets/uicrit"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:45Z", "role": "primary", "sha256": "8e790a53de8c40aa14f9e49aebf42b077e6a172a6841f38c7c6851af57b0617a", "supports": ["release_status"], "url": "https://api.github.com/repos/google-research-datasets/uicrit/contents"}], "id": "github:google-research-datasets/uicrit", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:02:45Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research-datasets/uicrit"}], "name": "UICrit", "notes": "The published CSV contains critique text, normalized boxes, ratings and RICO IDs; screenshots are referenced through RICO. The README explicitly licenses the work CC-BY-4.0. Its current 1,000-screen, 11,344-critique release differs from the original paper population. The README cites an ACM DOI rather than an arXiv ID, so no paper relation is asserted here. Its README establishes mobile interfaces without naming an operating system.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Mobile-interface critiques and quality ratings with region coordinates and RICO screen references, including a larger three-annotator release than the original paper dataset.", "topics": ["screen-understanding", "ui-design", "evaluation"], "updated_at": null, "url": "https://github.com/google-research-datasets/uicrit"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["Android device interactions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md", "population": "parameterized AndroidWorld tasks", "qualifier": "exact", "unit": "task definitions", "value": 116}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md", "population": "documented AndroidWorld app set", "qualifier": "exact", "unit": "applications", "value": 20}], "evaluation": "Programmatic task-specific initialization, system-state success checks, and teardown.", "evidence_urls": ["https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"], "focus": ["mobile task completion"], "notes": "Random parameters produce many instances from the same task definitions. Preserve seeds and parameterization; 116 is not a count of all possible task instances.", "observations": ["screenshots", "accessibility information"], "setting": "interactive"}, "evaluation_type": "programmatic state checks", "leaderboard_url": "https://docs.google.com/spreadsheets/d/1cchzP9dlTZ3WXQTfYNhh3avxoLipqHN75v1Tb86uhHo/edit?gid=0#gid=0", "metrics": ["task success"], "task_count": 116, "task_unit": "parameterized task definitions"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "93201de5e21f2b12283a75b1e2c357781b2e6cc27eab20cc7dd7aa0f5919d839", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.task_count", "details.task_unit", "details.metrics", "details.evaluation_type", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "b09fcaf4fbc265223c2479c410beef2b98c5e687c78ecedf3452aa12ffc465f1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/LICENSE"}], "id": "github:google-research/android_world", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research/android_world"}], "name": "AndroidWorld", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md", "target_id": "arxiv:2405.14573", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Android environment with programmatic task evaluation for mobile agents.", "topics": ["evaluation", "mobile-navigation"], "updated_at": null, "url": "https://github.com/google-research/android_world"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "63349e7557bd99240d4628de3994a53b69e54e3bd809895be11e9879bfba40cb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/gydpku/PPTC/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "f872ff2b1a064f46a45eee86f2cb2562fc0513285b80036c6cf7df9a5ffc6c70", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/gydpku/PPTC/HEAD/LICENSE"}], "id": "github:gydpku/pptc", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/gydpku/PPTC/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/gydpku/PPTC"}], "name": "PPTC", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/gydpku/PPTC/HEAD/README.md", "target_id": "arxiv:2311.01767", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "PowerPoint creation and editing benchmark with API-based execution and checks of the resulting presentation files.", "topics": ["evaluation", "tool-integration"], "updated_at": null, "url": "https://github.com/gydpku/PPTC"}
{"aliases": [], "authors": [], "details": {"availability": "Public benchmark code and public Zenodo dataset release metadata.", "benchmark_protocol": {"actions": ["generate procedures", "segment demonstrations", "answer workflow questions", "validate and rank procedures"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/HazyResearch/wonderbread/HEAD/README.md", "population": "WONDERBREAD BPM benchmark", "qualifier": "exact", "unit": "task types", "value": 6}, {"basis": "reported", "evidence_url": "https://zenodo.org/api/records/14094162", "population": "full dataset release 14094162", "qualifier": "exact", "unit": "human demonstrations", "value": 2928}, {"basis": "reported", "evidence_url": "https://zenodo.org/api/records/14094162", "population": "full dataset release 14094162", "qualifier": "exact", "unit": "web navigation workflows", "value": 598}, {"basis": "reported", "evidence_url": "https://zenodo.org/api/records/14094162", "population": "gold subset used for the original paper evaluations", "qualifier": "exact", "unit": "human demonstrations", "value": 724}], "evaluation": "Task-specific scripts evaluate six workflow-understanding task types; model outputs are not treated as successful interactive execution.", "evidence_urls": ["https://raw.githubusercontent.com/HazyResearch/wonderbread/HEAD/README.md", "https://zenodo.org/api/records/14094162"], "focus": ["workflow documentation", "workflow quality assessment", "workflow knowledge transfer"], "notes": "Counts are publisher-reported. The gold subset covers 162 workflows and differs from the full collection; scores require the exact task and subset.", "observations": ["screen recordings", "action traces", "screenshots", "written standard operating procedures"], "setting": "offline"}, "dataset_release": "10.5281/zenodo.14094162", "license_review": {"data": {"evidence_urls": ["https://zenodo.org/api/records/14094162"], "reason": "The linked Zenodo dataset release declares CC-BY-4.0; this does not assign a code license to the repository."}}, "repository_revision": "ed052c67aeada04167cdfe92ff8de454aa94627a"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:11Z", "role": "primary", "sha256": "1e4308d59a22b39543c28b745a58d3d7c47be1ab9df9a4c5c0fe868a810350b2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/HazyResearch/wonderbread/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:11Z", "role": "primary", "sha256": "5ba27e24455b36b52bf65b3c9f589e760f97a576bb01d622d89836c5bb53b38c", "supports": ["identity"], "url": "https://api.github.com/repos/HazyResearch/wonderbread"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:04Z", "role": "primary", "sha256": "accb518a5362948cf96e9f0b684fb6fcc9b26e43a691016ae70a3283c0e1e2c2", "supports": ["release_status", "licenses.code"], "url": "https://api.github.com/repos/HazyResearch/wonderbread/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:34Z", "role": "primary", "sha256": "68b4b952c1ab90484d457b392de1d0c95a24f6cc1674c82125a6bd26f1cb1016", "supports": ["release_status", "licenses.data", "details.benchmark_protocol"], "url": "https://zenodo.org/api/records/14094162"}], "id": "github:hazyresearch/wonderbread", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T09:12:11Z", "licenses": {"code": null, "data": {"basis": "The linked Zenodo dataset release declares CC-BY-4.0; this does not assign a code license to the repository.", "declared": "cc-by-4.0", "retrieved_at": "2026-10-17T09:15:34Z", "url": "https://zenodo.org/api/records/14094162"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/HazyResearch/wonderbread"}], "name": "WONDERBREAD", "notes": "The repository contains evaluation and collection code; the linked Zenodo release exposes the full, gold and debug demonstration archives and metadata. This measures workflow understanding rather than interactive agent task completion. Zenodo declares CC-BY-4.0 for that dataset; no repository-level code license was found. Archives and human demonstration records were not downloaded.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/HazyResearch/wonderbread/HEAD/README.md", "target_id": "arxiv:2406.13264", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An offline benchmark of workflow documentation, quality assessment and knowledge transfer using human web-navigation demonstrations, action traces and written procedures.", "topics": ["evaluation", "workflow", "video", "screen-understanding"], "updated_at": null, "url": "https://github.com/HazyResearch/wonderbread"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "68cb0deb8a2eb49fb9236ae3d885b043d240dded0fce31cf7425130a56ff5819", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/HeimingX/TAG/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "a33bee17685d64d23fb4d65b30b9c52f3a6dd5b21789d7b3509eb2e15145de6f", "supports": ["identity"], "url": "https://api.github.com/repos/HeimingX/TAG"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "fac4f5e38011bbcfef916063d495c21d49830759af4e8713a60d7bae992d2710", "supports": ["release_status"], "url": "https://api.github.com/repos/HeimingX/TAG/git/trees/main?recursive=1"}], "id": "github:heimingx/tag", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/HeimingX/TAG"}], "name": "TAG attention-driven GUI grounding", "notes": "The release contains grounding and benchmark evaluation code. The inspected root repository metadata does not declare a code license. It uses external pretrained models rather than publishing a new trained checkpoint, and the README restricts redistribution of unzipped benchmark data.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/HeimingX/TAG/HEAD/README.md", "target_id": "arxiv:2412.10840", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Grounding and evaluation code that aggregates a pretrained vision-language model’s attention maps to localize requested interface elements without fine-tuning.", "topics": ["grounding", "evaluation", "training-free"], "updated_at": null, "url": "https://github.com/HeimingX/TAG"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["PyAutoGUI", "UiPath Computer_13 action interface"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md", "population": "WindowsWorld benchmark described in the retained README", "qualifier": "exact", "unit": "tasks", "value": 181}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md", "population": "WindowsWorld coverage described in the retained README", "qualifier": "exact", "unit": "desktop applications", "value": 17}], "evaluation": "Publisher distinguishes S_int (intermediate checkpoint score on L1–L3) and S_final (final completion on L1–L4). Documented step budgets are 15/25/40/20 for L1/L2/L3/L4. Preserve level, observation mode, agent interface and checkpoint/final metric when reporting a run.", "evidence_urls": ["https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md", "https://api.github.com/repos/HITsz-TMG/WindowsWorld/git/trees/main?recursive=1"], "focus": ["multi-application Windows workflows", "intermediate-state and final-goal completion"], "notes": "Interactive VMware/Windows environment. The public repository includes benchmark.json, task/evaluator code and monitoring support; the README refers VM import to Installation Guide.md, whose distribution links and payloads were not inspected for this profile. Checkpoint counts and application counts are not extra tasks. Apache-2.0 is recorded for repository code; Windows, apps and VM artifacts retain separate terms.", "observations": ["screenshot", "accessibility tree", "screenshot plus accessibility tree", "set-of-marks"], "setting": "interactive"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "9857bd8ae9f521f152761ae3bd85debe4e7471aad34c9eb47618332948bca6c9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "490c9257d56f0a5214e6424e02490379b0b00ca2f7d4626fd5d14e17239d05ed", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/HITsz-TMG/WindowsWorld/git/trees/main?recursive=1"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "a8d1d88a0f9cce3cdbff73b03a80dc5e5d69e3ba2443946c268815a550c9e0ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/LICENSE"}], "id": "github:hitsz-tmg/windowsworld", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/HITsz-TMG/WindowsWorld"}], "name": "WindowsWorld", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md", "target_id": "arxiv:2604.27776", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Cross-application Windows benchmark with intermediate task evaluation.", "topics": ["evaluation", "cross-app"], "updated_at": null, "url": "https://github.com/HITsz-TMG/WindowsWorld"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "3ef331612819a006d4279ac59203971a22a56fb6d9a0047f85a82668d09c327b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/hud-evals/hud-python/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c88cc185e5547fce850b56f3b4072bed161658dc7277707e509437b9d760d470", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/hud-evals/hud-python/HEAD/LICENSE"}], "id": "github:hud-evals/hud-python", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/hud-evals/hud-python/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/hud-evals/hud-python"}], "name": "HUD", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "SDK for RL environments and evals spanning coding, browser, computer-use and robotics agents, whose environments expose SSH, MCP, CDP or VNC capabilities behind a minimal start-and-grade protocol.", "topics": ["evaluation", "reinforcement-learning"], "updated_at": null, "url": "https://github.com/hud-evals/hud-python"}
{"aliases": [], "authors": [], "details": {"availability": "Public benchmark-adapter source and running instructions.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/huggingface/screensuite/HEAD/LICENSE"], "reason": "Inspected repository LICENSE; not a license for linked benchmark datasets, pretrained weights or services."}}, "repository_revision": "9ceda5070d5664dd9376d5f8e7b43f02188e01ff"}, "environments": ["web", "android", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:13Z", "role": "primary", "sha256": "edcaa6f8355be6a8774ebeeb9bdb67648e8eeddb813d6be0fa2433116d9def36", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/huggingface/screensuite/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:12Z", "role": "primary", "sha256": "e372458e48d0ebbbe0f7887b66e2d5bb4db612fd96db4dc53404c0319891f7e8", "supports": ["identity"], "url": "https://api.github.com/repos/huggingface/screensuite"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:05Z", "role": "primary", "sha256": "93b685a13bfba4d9569dbfc0418de93c7b9f8a3f7c101af053aadeebbdf8b2c0", "supports": ["release_status"], "url": "https://api.github.com/repos/huggingface/screensuite/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:06Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/huggingface/screensuite/HEAD/LICENSE"}], "id": "github:huggingface/screensuite", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T09:12:13Z", "licenses": {"code": {"basis": "Inspected repository LICENSE; not a license for linked benchmark datasets, pretrained weights or services.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T09:14:06Z", "url": "https://raw.githubusercontent.com/huggingface/screensuite/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/huggingface/screensuite"}], "name": "ScreenSuite", "notes": "The publisher states that the suite compares the underlying multimodal models rather than full agent implementations. Public source includes benchmark adapters and registry code. Multi-step environments require Docker and KVM; constituent datasets, remote model services and OSWorld account-dependent tasks retain their own setup and terms. Apache-2.0 is scoped to this repository code.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A shared evaluation harness for GUI-model perception, offline action prediction and multi-step interaction, using simple smolagents-based agents across existing benchmarks.", "topics": ["evaluation", "grounding", "action-prediction"], "updated_at": null, "url": "https://github.com/huggingface/screensuite"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "9fe74ee087156efde4531a058ccd3abad504032e0bd994979992ed4164c8704b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/hyperbrowserai/HyperAgent/HEAD/README.md"}], "id": "github:hyperbrowserai/hyperagent", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/hyperbrowserai/HyperAgent"}], "name": "HyperAgent", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/hyperbrowserai/HyperAgent/HEAD/README.md", "target_id": "organization:hyperbrowser", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Library adding natural-language commands to Playwright: accessibility-tree single actions, multi-step tasks that can use screenshots, an MCP client and action caching for replay without LLM calls.", "topics": ["web-navigation"], "updated_at": null, "url": "https://github.com/hyperbrowserai/HyperAgent"}
{"aliases": [], "authors": [], "details": {"leaderboard_url": "https://github.com/iMeanAI/WebCanvas/blob/main/Mind2web-live_Leaderboard.md"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "5b2b4b4ce1241db7fd776f2ba0856ec2e0e441a56fa6461234abe0c6516f643e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/iMeanAI/WebCanvas/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "cbaab6bcbc9c3e28d0de902e12f92c2466103c775cb2a5af1e6d4110e82223a6", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/iMeanAI/WebCanvas/HEAD/LICENSE"}], "id": "github:imeanai/webcanvas", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/iMeanAI/WebCanvas/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/iMeanAI/WebCanvas"}], "name": "WebCanvas", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/iMeanAI/WebCanvas/HEAD/README.md", "target_id": "arxiv:2406.12373", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Framework and evaluation resources for real-world web agents.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/iMeanAI/WebCanvas"}
{"aliases": [], "authors": [], "details": {"availability": "Public CoAT prompting/action-prediction and evaluation source.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/IMNearth/CoAT/HEAD/README.md", "https://api.github.com/repos/IMNearth/CoAT/git/trees/main?recursive=1"], "reason": "Apache-2.0 badge alone does not establish component-scoped source terms; the inspected non-truncated tree has no license file."}}, "repository_revision": "9ab1a9f0e76581c50c84f8ebc6ada1a2084a0c71"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "7a9fb905b64f2760a137cd267a5e38dd343df1821eda8172b606ab7500d27731", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/IMNearth/CoAT/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "d0c5056c9a74913f84a4f809bf01d9fd9ea325ac0a91b4c4a0a51220b2da9d68", "supports": ["identity"], "url": "https://api.github.com/repos/IMNearth/CoAT"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:04Z", "role": "primary", "sha256": "29d84ed12dc25035fd539604b06f6cc113080b389edeb9aa605cfa55c5f28151", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/IMNearth/CoAT/git/trees/main?recursive=1"}], "id": "github:imnearth/coat", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:53Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/IMNearth/CoAT"}], "name": "CoAT research implementation", "notes": "The tree contains run_coat.py and coat/agent.py, model.py and evaluate.py, with API-backed model configuration. This verifies a released annotation/prediction implementation, not a live phone controller or a released checkpoint. AiTZ screens/annotations are linked through external Drive/Baidu bundles whose complete download inventory was not verified here. The README has an Apache-2.0 badge, but no license file or explicit component grant is present; code/data/weight rights remain unestablished.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/IMNearth/CoAT/HEAD/README.md", "target_id": "arxiv:2403.02713", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Research scripts that generate screen descriptions, action reasoning and expected outcomes, then predict Android GUI actions using the resulting Chain-of-Action-Thought context.", "topics": ["planning", "mobile-navigation", "action-prediction"], "updated_at": null, "url": "https://github.com/IMNearth/CoAT"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "5703c892aaf99d250bf59edf52d359560a80d443f2be13c7aa76af0c1a7f1d67", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md"}], "id": "github:inclusionai/ui-venus", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/inclusionAI/UI-Venus"}], "name": "UI-Venus", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md", "target_id": "arxiv:2508.10833", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md", "target_id": "arxiv:2602.09082", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md", "target_id": "arxiv:2609.00028", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI agent model family; UI-Venus-2, at 9B and 27B, acts on mobile, web and desktop in one end-to-end agent, trained with keypoint-based multi-judge task verification.", "topics": ["grounding", "planning"], "updated_at": null, "url": "https://github.com/inclusionAI/UI-Venus"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "4b68c82a2dd0063a450bde34c8bffbd931565a91a2982a5b47be7e711815b284", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/LICENSE"}], "id": "github:infixai/infigui-g1", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/InfiXAI/InfiGUI-G1"}], "name": "InfiGUI-G1", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/README.md", "target_id": "arxiv:2508.05731", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "InfiGUI-G1 3B and 7B grounding models trained with Adaptive Exploration Policy Optimization, which samples several candidate answers per pass under an adaptive exploration reward.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/InfiXAI/InfiGUI-G1"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:51Z", "role": "primary", "sha256": "dfcbaa088cc21dd4e3c67d6eb3b3a1dc476217165ce2ca0daa2facfda71b5a38", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://github.com/jacklishufan/MobileWorld"}], "id": "github:jacklishufan/mobileworld", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/jacklishufan/MobileWorld"}], "name": "MobileWorldBench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://github.com/jacklishufan/MobileWorld", "target_id": "arxiv:2512.14014", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Benchmark of semantic world modeling for mobile agents, with question-answering tasks and GPT-judged generation tasks, plus training data built on AitW and AndroidControl screens.", "topics": ["evaluation", "mobile-navigation"], "updated_at": null, "url": "https://github.com/jacklishufan/MobileWorld"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:05Z", "role": "primary", "sha256": "2316acbef1207925b10b4770e14bdf69f6f4b6b3bc0ddb12e6be107d6f1425da", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T01:52:51Z", "role": "primary", "sha256": "57bcc2eb985fa6036bf2bde7f28c4fd359c6b3479a658fea768162cac2a75ba5", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/LICENSE"}], "id": "github:jcwang0602/seg2click", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T01:46:05Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T01:52:51Z", "url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/jcwang0602/Seg2Click"}], "name": "Seg2Click", "notes": "The repository provides training, evaluation, and a custom mask decoder. Its code license does not determine the licenses of the five model checkpoints. No model execution was performed.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training and inference implementation that predicts an executable-region mask and derives a GUI click from that mask.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://github.com/jcwang0602/Seg2Click"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["GUI interactions", "command-line operations"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md", "population": "documented MacAgentBench suite", "qualifier": "exact", "unit": "tasks", "value": 676}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md", "population": "documented application coverage", "qualifier": "exact", "unit": "applications", "value": 25}], "evaluation": "Deterministic rule-based evaluators with intermediate checkpoints.", "evidence_urls": ["https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"], "focus": ["macOS task completion", "cross-application workflows"], "notes": "The benchmark evaluates multiple agent frameworks and GUI/CLI workflows. Framework skills and interfaces affect the task setting; the inspected README does not define one universal observation representation.", "observations": [], "setting": "interactive"}, "leaderboard_url": "https://jetastra.github.io/MacAgentBench/"}, "environments": ["desktop", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "f1f8c0cd5d984e4026abea858b54f0f1427c018cad7e75322e744f1e070225a9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["details.benchmark_protocol", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "adf69997028988ba1eec34666afa119e007771dfc8e830c5dc50ad650896e897", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/LICENSE"}], "id": "github:jetastra/macagentbench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/JetAstra/MacAgentBench"}], "name": "MacAgentBench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48", "target_id": "arxiv:2606.22557", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "676 macOS tasks across 25 applications, each run in its own Docker container and scored by deterministic rule-based checks with multiple checkpoints.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/JetAstra/MacAgentBench"}
{"aliases": ["github:dvlab-research/arpo", "dvlab-research/ARPO", "https://github.com/dvlab-research/ARPO"], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "7489ae309e90d8b3b9e823f495b63e5116434060", "repository_url": "https://github.com/JIA-Lab-research/ARPO"}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:12Z", "role": "primary", "sha256": "839b2de0e0c614fe8d2df4782d0f6a6e7d33f5c60d8020c9d4a01c61d533f39e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:39Z", "role": "primary", "sha256": "bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898", "supports": ["identity"], "url": "https://api.github.com/repos/JIA-Lab-research/ARPO"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:23Z", "role": "primary", "sha256": "b3ff7f4e350223827cae0fffd986ee672e0b4466349c262ed2dc9e63a8480aeb", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/JIA-Lab-research/ARPO/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:00Z", "role": "primary", "sha256": "bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898", "supports": ["identity"], "url": "https://api.github.com/repos/dvlab-research/ARPO"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T15:37:00Z", "role": "primary", "sha256": "bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898", "supports": ["aliases"], "url": "https://api.github.com/repos/dvlab-research/ARPO"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-18T04:40:39Z", "role": "primary", "sha256": "bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898", "supports": ["aliases"], "url": "https://api.github.com/repos/JIA-Lab-research/ARPO"}], "id": "github:jia-lab-research/arpo", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:12Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/JIA-Lab-research/ARPO"}], "name": "ARPO training framework", "notes": "Official GitHub API resolves the former dvlab-research/ARPO URL to JIA-Lab-research/ARPO with the same numeric repository ID. Training/evaluation scripts are released; OSWorld environments and the separate model checkpoint are not bundled into this code record. Reported benchmark scores were not reproduced.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/README.md", "target_id": "arxiv:2505.16282", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Replay-based policy-optimization framework for GUI agents, with distributed OSWorld rollouts and screenshot/action histories.", "topics": ["reinforcement-learning", "long-horizon", "automation"], "updated_at": null, "url": "https://github.com/JIA-Lab-research/ARPO"}
{"aliases": ["github:gimme1dollar/b-moca", "gimme1dollar/b-moca", "https://github.com/gimme1dollar/b-moca"], "authors": [], "details": {"benchmark_protocol": {"actions": ["normalized touch/lift coordinate pairs for tap or swipe"], "counts": [{"basis": "computed", "evidence_url": "https://api.github.com/repos/jylee425/b-moca/git/trees/release/ver.3?recursive=1", "population": "asset/tasks/**/*.textproto in retained release/ver.3 tree; not a paper test-task total", "qualifier": "exact", "unit": "task-definition files", "value": 131}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_train.csv", "population": "environments_train.csv, retained release/ver.3 snapshot", "qualifier": "exact", "unit": "environment configuration rows", "value": 35}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_test.csv", "population": "environments_test.csv, retained release/ver.3 snapshot", "qualifier": "exact", "unit": "environment configuration rows", "value": 10}], "evaluation": "Task success is detected using Appium; positive task reward terminates the episode. Environment variants change device/locale/wallpaper/DPI/dark mode and are loaded as snapshots. The numeric action interface is (touch_y, touch_x, lift_y, lift_x); wrapper/tanh transformations must be kept with the evaluated agent.", "evidence_urls": ["https://raw.githubusercontent.com/jylee425/b-moca/HEAD/readme.md", "https://api.github.com/repos/jylee425/b-moca/git/trees/release/ver.3?recursive=1", "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/bmoca/environment/coordinator.py", "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/bmoca/environment/environment.py", "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_train.csv", "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_test.csv"], "focus": ["Android control across visual/device variation", "generalization across environment snapshots"], "notes": "Current source inventory differs in grain from published experiment subsets: do not multiply 131 definitions by 35/10 configuration rows or label the product as evaluated tasks. Setup requires Android API 29, Appium, external app versions and generated AVD snapshots. No apps/snapshots were downloaded or run. Code Apache-2.0 does not establish data or app rights.", "observations": ["screenshot pixels", "XML view hierarchy", "goal instruction"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "bd1354c8008d7d79d4883c0c037b7918392f86fab61391b39b52e507b04672ce", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "06ba40a5e0d5fb52e7bf5e5df636cb8e8bbbc56a8308dc6d528012489f7650ee", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/jylee425/b-moca/git/trees/release/ver.3?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:24:53Z", "role": "primary", "sha256": "34bc0c23949732e0a8091e334d59d2d41cae49a82cde850247aa0e6d739d2464", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/bmoca/environment/coordinator.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "d866af344c0cff5255b813576f2e9e9c5c2964e6809058b39e2b5180af71d946", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/bmoca/environment/environment.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:23:14Z", "role": "primary", "sha256": "c278cc28f8a0667b7fdc6ba638704cfbe43783cd496181231c45c865087543f3", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_train.csv"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:23:15Z", "role": "primary", "sha256": "4b92933fe4f4f66e71a694b579940db11bedd5ea8f153752a44e633825b06fa3", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_test.csv"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:58Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:03Z", "role": "primary", "sha256": "9d4706306b04960f0eb4298887496961efdee6b7d41445e82abcac489d63779f", "supports": ["aliases"], "url": "https://api.github.com/repos/gimme1dollar/b-moca"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:04Z", "role": "primary", "sha256": "9d4706306b04960f0eb4298887496961efdee6b7d41445e82abcac489d63779f", "supports": ["aliases"], "url": "https://api.github.com/repos/jylee425/b-moca"}], "id": "github:jylee425/b-moca", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:58Z", "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/jylee425/b-moca"}], "name": "B-MoCA", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Mobile control benchmark with varied interface configurations.", "topics": ["evaluation", "robustness"], "updated_at": null, "url": "https://github.com/jylee425/b-moca"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["tap/long-tap/swipe", "navigation buttons and text editing", "refuse, ask-consent and complete", "task-specific shortcut actions"], "counts": [], "evaluation": "Appium-backed task evaluators track goal achievement and harm prevention as separate booleans, plus risk-detection step and completion progress. The base evaluator defaults to max_step=15; task-specific overrides must be retained. Results are not a single generic success-rate metric.", "evidence_urls": ["https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/README.md", "https://api.github.com/repos/jylee425/mobilesafetybench/git/trees/release/ver.3?recursive=1", "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/environment.py", "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/component/parser.py", "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/evaluator/_base_evaluator.py"], "focus": ["mobile task helpfulness", "harm prevention during Android interaction"], "notes": "Inspected release/ver.3 branch exposes task definitions, Android environment and per-task evaluators. It requires Android API 34, Appium/ADB and documented third-party app versions; those APKs/accounts were not acquired. The action parser includes semantic shortcuts in addition to screen gestures, so this is not a strictly coordinate-only action setting. No corpus task total is inferred from evaluator-module or file counts. Apache-2.0 applies to repository code, not automatically app/data assets.", "observations": ["screenshot pixels", "parsed UI hierarchy text", "task instruction and interaction history"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "e733115d9b2fbe910cb3891162f04a89af8cbda25255c8f0bf60f78101e4309a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "8dc66f6c5087a32c5c9fb7b3104ea6f9e0082255e2ba3cf9ba8244ce5a24da6f", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/jylee425/mobilesafetybench/git/trees/release/ver.3?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "f56d6345546a9c67a2113cfba5b7148f5bd6b48fac5c34f1a5f4080e2d81ae2f", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/environment.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "120487c3a6492371321dd3eebcd65dc59de6a82454d622d7fed3a7d86b8fa6c5", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/component/parser.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:23:14Z", "role": "primary", "sha256": "918e3aa64ab3f578fb54580da123c1ff4508022b64ce74f88ce28c7a612756a4", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/evaluator/_base_evaluator.py"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:58Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/LICENSE.txt"}], "id": "github:jylee425/mobilesafetybench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:58Z", "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/LICENSE.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/jylee425/mobilesafetybench"}], "name": "MobileSafetyBench", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Android emulator testbed that scores mobile device-control agents on both safety and helpfulness, using ADB and Appium for observations, actions and success checks.", "topics": ["safety-security", "evaluation"], "updated_at": null, "url": "https://github.com/jylee425/mobilesafetybench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "eb2a2767741b53e5a6659ef4b0cc325ac18ffdc4222deef1d0e7671089fcaa39", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "a145c4ec8fd43f59ebe97708b6ee4cef8112959036957eaa4739b8da9a297e8a", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/LICENSE"}], "id": "github:kdegroup/ui-agile", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/KDEGroup/UI-AGILE"}], "name": "UI-AGILE", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/README.md", "target_id": "arxiv:2507.22025", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI agent trained by RL with a continuous reward, a simple-thinking reward and cropping-based resampling on about 9k samples, plus a plug-in decomposed grounding-and-selection inference method.", "topics": ["grounding", "reinforcement-learning"], "updated_at": null, "url": "https://github.com/KDEGroup/UI-AGILE"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/LICENSE", "https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/README.md"], "reason": "The README declares Apache-2.0, but the inspected LICENSE is a short notice and link rather than the complete terms; the code license remains unconfirmed."}}, "repository_id": 1278022559}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:54:57.854138+00:00", "role": "primary", "sha256": "3f7545f466407ed5336f47d18b12e8354ec352ff83dcf973b1545f81d9a629ac", "supports": ["name", "kind", "summary", "topics", "scope", "relations", "notes", "licenses.code"], "url": "https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:55:26.237759+00:00", "role": "primary", "sha256": "4c92c584b6885a3bde8ee89d04ccf48287e7f6119889f75cde0c034c90bd8217", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/evaluation/memgui3k_offline_eval/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:55:01.061844+00:00", "role": "primary", "sha256": "d4e344274155be1d0a192c8ae2940169abb0bd8546d045b96b45810593860636", "supports": ["licenses.code", "notes"], "url": "https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-19T19:10:26.954787+00:00", "role": "primary", "sha256": "0ae4477b885b0915e43874399c2a6b91f7d07954697370204550df931f7db82c", "supports": ["details.repository_id"], "url": "https://api.github.com/search/repositories?q=topic%3Agui-agent+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-18&per_page=100&page=1&sort=updated&order=asc"}], "id": "github:kwai/memgui-agent", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-20T03:54:57.854138+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/kwai/MemGUI-Agent"}], "name": "MemGUI-Agent", "notes": "The inspected release documentation describes dataset utilities, a LoRA SFT template and gold-step offline evaluation; live Android rollouts are explicitly outside that evaluator and linked to separate benchmark repositories. Executable source files, external datasets/checkpoints and reported results were not independently verified in this bounded review.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/README.md", "target_id": "arxiv:2606.19926", "type": "introduced_by"}], "review_status": "curated", "scope": "core", "summary": "MemGUI-Agent training and offline evaluation tools provide trajectory conversion, LoRA SFT templates, and step-level checks of GUI actions, memory updates and context folding.", "topics": ["memory", "interaction-history", "long-horizon", "imitation-learning", "evaluation"], "updated_at": null, "url": "https://github.com/kwai/MemGUI-Agent"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/kwai/MobileForge/main/LICENSE", "https://raw.githubusercontent.com/kwai/MobileForge/main/README.md"], "reason": "The complete root LICENSE declares Apache-2.0 for repository code. Linked data and model weights, and third-party dependencies, retain separate licensing requirements."}}, "repository_id": 1278021888}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:55:04.230792+00:00", "role": "primary", "sha256": "f78f58a6e3b8dd2ae3e8c49a15473782ae258ecd70719f7eeddfa43a10441adf", "supports": ["name", "kind", "summary", "topics", "scope", "relations", "notes", "environments", "licenses.code"], "url": "https://raw.githubusercontent.com/kwai/MobileForge/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:55:29.533431+00:00", "role": "primary", "sha256": "58187fe25d8398ef2ae31de02debd4bc3b932ff1925db86a87f5de65b016aed1", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/kwai/MobileForge/main/training/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:55:07.420029+00:00", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code", "notes"], "url": "https://raw.githubusercontent.com/kwai/MobileForge/main/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-19T19:10:26.954787+00:00", "role": "primary", "sha256": "0ae4477b885b0915e43874399c2a6b91f7d07954697370204550df931f7db82c", "supports": ["details.repository_id"], "url": "https://api.github.com/search/repositories?q=topic%3Agui-agent+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-18&per_page=100&page=1&sort=updated&order=asc"}], "id": "github:kwai/mobileforge", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-20T03:55:04.230792+00:00", "licenses": {"code": {"basis": "Complete repository root Apache License 2.0 text; code declaration only, not linked datasets or weights.", "declared": "Apache-2.0", "retrieved_at": "2026-10-20T03:55:07.420029+00:00", "url": "https://raw.githubusercontent.com/kwai/MobileForge/main/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/kwai/MobileForge"}], "name": "MobileForge", "notes": "The root README describes MobileGym exploration/feedback and HiFPO optimization. The training README limits its public training path to the Qwen3-VL-8B experiment. Linked model and dataset collections, executable source completeness and benchmark results were not independently verified; the repository code license is not a model or dataset license.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/kwai/MobileForge/main/README.md", "target_id": "arxiv:2606.19930", "type": "introduced_by"}], "review_status": "curated", "scope": "core", "summary": "Android GUI-agent adaptation framework combining app exploration, generated curricula and rollout feedback with hint-conditioned step-level GRPO; the documented public training path targets Qwen3-VL-8B.", "topics": ["mobile-navigation", "data-generation", "reinforcement-learning", "evaluation"], "updated_at": null, "url": "https://github.com/kwai/MobileForge"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "07431058d8b9a71b80e75b7a74a40a8346c1503eec240628459cc4d4e2f0802f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "52925efd66489696c4e1f31f0f8353997fe534df9fcd42e50232cc78f0506903", "supports": ["identity"], "url": "https://api.github.com/repos/kyle8581/Web-Shepherd"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "ba7d28866cec4fb889499fc011473aedcdcc36fd8175c73527eafd7ef6b6dbba", "supports": ["release_status"], "url": "https://api.github.com/repos/kyle8581/Web-Shepherd/git/trees/main?recursive=1"}], "id": "github:kyle8581/web-shepherd", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T02:44:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/kyle8581/Web-Shepherd"}], "name": "Web-Shepherd inference and evaluation", "notes": "The tree contains src/webprm modules and a WebRewardBench evaluation script. The README’s illustrative folder layout differs from the actual tree, so a complete training pipeline is not claimed. No root code license is declared. Some old checkpoint links are unavailable; current model releases are separately identified through the official Hub collection.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Process-reward inference and benchmark-evaluation utilities that use task checklists to assess individual web-agent decisions.", "topics": ["reward-models", "evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/kyle8581/Web-Shepherd"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "9b4b945f4514d261373ec08d4c3e4e86b6623f95f95b9469bec2b8e7e93d9807", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/LaoKuiZe/AppAgent-Pro/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "f13a0d1aaaec0f7e32f034197b31b2bc0537685aff11af653381454288c75794", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/LaoKuiZe/AppAgent-Pro/HEAD/LICENSE"}], "id": "github:laokuize/appagent-pro", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/LaoKuiZe/AppAgent-Pro/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/LaoKuiZe/AppAgent-Pro"}], "name": "AppAgent Pro", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/LaoKuiZe/AppAgent-Pro/HEAD/README.md", "target_id": "arxiv:2508.18689", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Proactive Android agent built on AppAgent that uses GPT-4o to decide whether YouTube or Amazon content would help, operates those apps over ADB, and merges what it finds into its answer.", "topics": ["mobile-navigation"], "updated_at": null, "url": "https://github.com/LaoKuiZe/AppAgent-Pro"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "7e4d38b95657b604f5f390e25762d94c8d59164fbb2ae6c140c126bb128dbd4f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/lavague-ai/LaVague/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/lavague-ai/LaVague/HEAD/LICENSE"}], "id": "github:lavague-ai/lavague", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/lavague-ai/LaVague/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/lavague-ai/LaVague"}], "name": "LaVague", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Open-source web agent framework where a World Model turns an objective and the current page into instructions and an Action Engine compiles them into Selenium or Playwright code to run.", "topics": ["web-navigation"], "updated_at": null, "url": "https://github.com/lavague-ai/LaVague"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:56Z", "role": "primary", "sha256": "95184255b6ec021aac8ca062b944f3ed76d1e01405e5771d9e31368021523eec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "notes"], "url": "https://raw.githubusercontent.com/lgy0404/LearnAct/HEAD/README.md"}], "id": "github:lgy0404/learnact", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-18T10:59:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/lgy0404/LearnAct"}], "name": "LearnAct", "notes": "At the retained README snapshot, LearnGUI-Online code and the real-world deployment implementation were marked coming soon. The mobile operating system is not established by the reviewed release instructions; no code was executed.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/lgy0404/LearnAct/HEAD/README.md", "target_id": "arxiv:2504.13805", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Released inference and evaluation code for demonstration-conditioned mobile GUI agents on LearnGUI-Offline. The repository describes a multi-agent adaptation method and links the LearnGUI demonstrations.", "topics": ["mobile-navigation", "imitation-learning", "evaluation"], "updated_at": null, "url": "https://github.com/lgy0404/LearnAct"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["predicted target coordinates"], "counts": [{"basis": "computed", "evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "population": "sum of icon/text counts in the published card table", "qualifier": "exact", "unit": "target annotations", "value": 1581}], "evaluation": "For positive targets, the predicted point must fall inside the annotated bounding box.", "evidence_urls": ["https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/readme.md", "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py", "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"], "focus": ["GUI grounding"], "notes": "Language, instruction style, and positive/negative target configuration are explicit evaluator options. This profile describes target localization, not end-to-end workflow completion. The target count is computed from the card table, not from expanded language/style evaluation instances.", "observations": ["high-resolution screenshots", "target instructions"], "setting": "offline"}, "leaderboard_url": "https://gui-agent.github.io/grounding-leaderboard"}, "environments": ["desktop", "windows", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "7b64ccad9cfec24c2c9cb35f55923bff93d53729cef3cc7463c79b29bdb1de1a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:38:05Z", "role": "primary", "sha256": "a77c859775b33591b5619e8520c94702a53a1f9a29ada93e66744dbe23afe9dc", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:35:55Z", "role": "primary", "sha256": "76ae2231564fe84af029f8625c02d9cd7fb6cf85c5f8af1b2a2593e2107719d9", "supports": ["details.benchmark_protocol", "environments"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "c78681e624c748b7a9980b6a9fd1514cec5140c7b3c7f08fe21b36bc7468f2ce", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/LICENSE"}], "id": "github:likaixin2000/screenspot-pro-gui-grounding", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding"}], "name": "ScreenSpot Pro", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI grounding benchmark focused on professional high-resolution interfaces.", "topics": ["evaluation", "grounding"], "updated_at": null, "url": "https://github.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["touch actions", "MCP tools (optional)"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md", "population": "tasks.json benchmark", "qualifier": "exact", "unit": "tasks", "value": 133}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md", "population": "bundled SwiftUI apps", "qualifier": "exact", "unit": "applications", "value": 26}], "evaluation": "Rubric satisfaction per task; full-task pass requires every criterion to be satisfied.", "evidence_urls": ["https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md"], "focus": ["personalized phone tasks", "cross-app workflows"], "notes": "Vision-only, vision-plus-XML, and MCP modes expose different information and tools. The documented default evaluator is model-based; compare the same mode and judge.", "observations": ["screenshots", "accessibility XML (optional)"], "setting": "interactive"}, "evaluation_type": "rubric-based", "metrics": ["mean rubric score", "full-task pass rate"], "task_count": 133}, "environments": ["ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "706f9d770b65a1238e56f74f7dd5089c7756cc88e0ac09d80c7b2b3085a09b27", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.task_count", "details.metrics", "details.evaluation_type", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "fa9fe1e31ed8da21d0b51621d3a3669ced4f0e42353aaf4bd8be1615d200eaac", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/LICENSE"}], "id": "github:ljang0/iosworld", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ljang0/iOSWorld"}], "name": "iOSWorld", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Native iPhone-simulator benchmark with shared user data, application workflows, agent runners, and rubric-based scoring.", "topics": ["evaluation", "cross-app", "memory"], "updated_at": null, "url": "https://github.com/ljang0/iOSWorld"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["device actions through AgentEnv"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md", "population": "complete task collection", "qualifier": "exact", "unit": "tasks", "value": 495}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md", "population": "AITW-derived subset", "qualifier": "exact", "unit": "tasks", "value": 102}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md", "population": "self-constructed subset", "qualifier": "exact", "unit": "tasks", "value": 393}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md", "population": "self-constructed subset", "qualifier": "exact", "unit": "Android applications", "value": 46}], "evaluation": "Checks whether the recorded execution visits all annotated essential states rather than requiring an identical action sequence.", "evidence_urls": ["https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md"], "focus": ["mobile task completion", "essential-state evaluation"], "notes": "Application coverage refers to the self-constructed subset. Task execution and essential-state annotation are separate components; counts are publisher-reported, not a fresh task-file recount.", "observations": ["screenshots", "view hierarchies", "Android activity names"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "4c4fe0fa92961f7286b6553b13117b2942e8482a48b7120a79e57dce2f8d42b9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "4c4fe0fa92961f7286b6553b13117b2942e8482a48b7120a79e57dce2f8d42b9", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md"}], "id": "github:llamatouch/llamatouch", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/LlamaTouch/LlamaTouch"}], "name": "LlamaTouch", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md", "target_id": "arxiv:2404.16054", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile UI task evaluation using application state information.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/LlamaTouch/LlamaTouch"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f55ac6a0c50d298d46c5f7083a0879959fd7cd3c1d91496af295ebf2afcc13a3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/LICENSE"}], "id": "github:lll6gg/ui-r1", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/lll6gg/UI-R1"}], "name": "UI-R1", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md", "target_id": "arxiv:2503.21620", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Rule-based reinforcement learning for GUI action prediction, releasing UI-R1-3B, trained from Qwen2.5-VL-3B, the efficiency-oriented UI-R1-E-3B, and training data.", "topics": ["reinforcement-learning", "grounding"], "updated_at": null, "url": "https://github.com/lll6gg/UI-R1"}
{"aliases": [], "authors": [], "details": {"availability": "Public memory, prompting, browser-experiment and fine-tuning source.", "license_conflict": "README MIT badge versus NTUITIVE agreement; the agreement contains inconsistent distribution provisions.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ltzheng/synapse/HEAD/README.md", "https://raw.githubusercontent.com/ltzheng/Synapse/HEAD/LICENSE"], "reason": "Unresolved conflict between README MIT badge and NTUITIVE non-commercial LICENSE, including inconsistent redistribution clauses. Keep permissions unknown rather than choose an interpretation."}}, "repository_revision": "08c3a252ff158822e50e9c109c1cd49dc2f6147d"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "ed5416d91665017fca4b354016b7c2c748acf32d7723483fecb5387a22a1fdff", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.code"], "url": "https://raw.githubusercontent.com/ltzheng/synapse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "79a238a6f486d2a8970baba199f0968d28f14d42922c86ce71753eb687e0e9c6", "supports": ["identity"], "url": "https://api.github.com/repos/ltzheng/synapse"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "785a7163dac9f41b1d0ca97f8dd9311b6df6a09f3a5163438dd64db226a4ceaf", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/ltzheng/Synapse/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:09Z", "role": "primary", "sha256": "cc4bce2274289d2b924fcc3c185c68b8e7c539365c5f8c72be865883244088ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ltzheng/Synapse/HEAD/LICENSE"}], "id": "github:ltzheng/synapse", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ltzheng/Synapse"}], "name": "Synapse", "notes": "Training, memory-building and evaluation scripts for MiniWoB++ and Mind2Web are present. External benchmark data, model/API credentials and legacy dependencies are still required. License remains unknown: the README shows MIT, while LICENSE is an NTUITIVE non-commercial agreement that both permits some distribution and separately says distribution is prohibited. No interpretation resolves those contradictions, and no commercial or non-commercial permission is asserted. This admission confirms the released GUI implementation, not a licensing grant.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50", "target_id": "arxiv:2306.07863", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A computer-control research implementation that retrieves example trajectories from memory, abstracts task-relevant state and prompts an agent to navigate browser tasks.", "topics": ["memory", "planning", "web-navigation", "imitation-learning"], "updated_at": null, "url": "https://github.com/ltzheng/Synapse"}
{"aliases": [], "authors": [], "details": {"availability": "Public graph/task manifests, evaluation source and ungated screenshot archive manifest.", "benchmark_protocol": {"actions": ["mobile GUI operations along graph transitions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md", "population": "reported ColorBench benchmark", "qualifier": "exact", "unit": "tasks", "value": 175}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md", "population": "reported ColorBench benchmark", "qualifier": "exact", "unit": "cross-app tasks", "value": 101}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md", "population": "reported ColorBench benchmark", "qualifier": "exact", "unit": "single-app tasks", "value": 74}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md", "population": "reported ColorBench benchmark", "qualifier": "exact", "unit": "mobile applications", "value": 21}], "evaluation": "The publisher reports success rate, completion rate and atomic-capability diagnostics using graph-based evaluation.", "evidence_urls": ["https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md"], "focus": ["long-horizon mobile task completion", "multi-path graph interaction"], "notes": "The interaction environment replays graph states rather than running real user accounts. The 101 and 74 subsets partition the reported 175 tasks; do not add the three counts.", "observations": ["mobile screenshots represented as graph states"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/LICENSE"], "reason": "Repository source LICENSE."}, "data": {"evidence_urls": ["https://huggingface.co/datasets/Norsheep/ColorBench/raw/main/README.md"], "reason": "Publisher metadata for the separately hosted screenshot archive; underlying third-party content retains applicable rights."}}, "repository_revision": "1f26a84230d26f99d84eab97d05032582b39018e"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:16Z", "role": "primary", "sha256": "5c0190fffb68b91c1855a408f6a5e252871d94000b95fa258504a3c100600cec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:16Z", "role": "primary", "sha256": "2f6940503cc0d6cb97bdf142d6b8c6ab116de82227f2fa6d97fd56d7e5aebec0", "supports": ["identity"], "url": "https://api.github.com/repos/MadeAgents/ColorBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:54Z", "role": "primary", "sha256": "6a00c093a6794ec4a172d30b70fdabc17400584a9e25a157a47a7661dbdd20fa", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/MadeAgents/ColorBench/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:06Z", "role": "primary", "sha256": "c0c3286b5150e128e55034ba0393ef85f7fcf73e4b73f4276389a04c9bcec919", "supports": ["identity", "release_status"], "url": "https://huggingface.co/api/datasets/Norsheep/ColorBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:06Z", "role": "primary", "sha256": "4bcf87ecfbbb8e07a01b21415a970c8b53a5283bf6872b657040d3f45c9241f7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/Norsheep/ColorBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:54Z", "role": "primary", "sha256": "130976c026bcafa9dcc479697fe6595dbefb73cdb375fc47b8478acc16e6ea73", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/LICENSE"}], "id": "github:madeagents/colorbench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T11:18:16Z", "licenses": {"code": {"basis": "Repository source LICENSE.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T11:22:54Z", "url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/LICENSE"}, "data": {"basis": "Publisher metadata for the separately hosted screenshot archive; underlying third-party content retains applicable rights.", "declared": "apache-2.0", "retrieved_at": "2026-10-17T11:23:06Z", "url": "https://huggingface.co/datasets/Norsheep/ColorBench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MadeAgents/ColorBench"}], "name": "ColorBench", "notes": "The source tree contains graph and task JSON files plus construction/evaluation code; the linked ungated Hub provides the screenshot archive. The environment is static but interactive and graph-based, not a live-device deployment. Counts are publisher-reported task/app inventories. Apache-2.0 declarations for repository code and the Hub dataset are separately retained; third-party application visuals are not granted additional rights by this catalog.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md", "target_id": "arxiv:2510.14621", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A graph-based mobile GUI benchmark that evaluates long tasks through screenshot states and multiple valid or erroneous transitions, including reflection, replanning and backtracking.", "topics": ["evaluation", "mobile-navigation", "long-horizon", "cross-app"], "updated_at": null, "url": "https://github.com/MadeAgents/ColorBench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "101595fa1f62d30271b0560f8aaa54286d7f6fc006e2aa9d8b73086583b27656", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MadeAgents/mobile-use/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "806f2d2c84a015f221f39691c95f7bdd0ff18f134e563d1155caa787076653fe", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/MadeAgents/mobile-use/HEAD/LICENSE"}], "id": "github:madeagents/mobile-use", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/MadeAgents/mobile-use/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MadeAgents/mobile-use"}], "name": "Mobile Use", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/MadeAgents/mobile-use/HEAD/README.md", "target_id": "arxiv:2507.16853", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Android GUI agent that operates real phones through ADB with hierarchical reflection and a multi-agent framework, run from a web interface or Python and evaluated on AndroidWorld and AndroidLab.", "topics": ["mobile-navigation"], "updated_at": null, "url": "https://github.com/MadeAgents/mobile-use"}
{"aliases": [], "authors": [], "details": {"availability": "Public conversational planning, candidate-generation and training/evaluation source.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/magicgh/Self-MAP/HEAD/LICENSE"], "reason": "Project source LICENSE; separate dataset license is not folded into the code slot."}}, "repository_revision": "6cf779682b70dae597626f96dfa5130b202689bc"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "693197cb807e3ab4d5afda9de35dd51342b4feb18e68b8ccfcdaa370399dbeb3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/magicgh/self-map/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "82c425841c94bf129c4fbd1e70029c581aa1488d5f85cdd23aa0f292092146f0", "supports": ["identity"], "url": "https://api.github.com/repos/magicgh/self-map"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "8d844f227c457321640231ab3a4953457e3c6cfeab8d8f4803f8aa01dc15a8f8", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/magicgh/Self-MAP/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "f271c479b19a620600d18fc2265224205bf71483b969672d3f6acd9bc552b3f5", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/magicgh/Self-MAP/HEAD/LICENSE"}], "id": "github:magicgh/self-map", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:55Z", "licenses": {"code": {"basis": "Project source LICENSE; separate dataset license is not folded into the code slot.", "declared": "MIT", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/magicgh/Self-MAP/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/magicgh/Self-MAP"}], "name": "Self-MAP", "notes": "The source tree provides candidate-generation and action-prediction training/evaluation scripts. Its documented workflow requires separately downloaded MT-Mind2Web data and configured model/API paths. The README cites the ACL publication by DOI; a retained original arXiv abstract explicitly names Self-MAP and MT-Mind2Web and supplies the catalog paper relation without a new bibliography entity. MIT is the code license; the separate data artifact declares CC-BY-4.0.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2402.15057", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/magicgh/self-map/HEAD/README.md", "target_id": "url:bdfb50330deb97893ce4", "type": "uses"}], "review_status": "curated", "scope": "core", "summary": "A conversational web-agent implementation with candidate ranking and memory-assisted action planning, including optional memory simplification, refinement and self-reflection modules.", "topics": ["web-navigation", "memory", "self-reflection", "planning"], "updated_at": null, "url": "https://github.com/magicgh/Self-MAP"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["judge verdicts"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending", "population": "paper-reported expert-reviewed set", "qualifier": "exact", "unit": "trajectories", "value": 1302}], "evaluation": "Compare automatic judgments with expert answers on task success, side effects and repetitiveness.", "evidence_urls": ["https://raw.githubusercontent.com/McGill-NLP/agent-reward-bench/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"], "focus": ["automatic evaluator agreement", "task success", "side effects", "repetition"], "notes": "Counts refer to the paper study; current Hub package files are not converted into trajectory totals.", "observations": ["web trajectories", "screenshots", "actions"], "setting": "offline"}, "leaderboard_url": "https://huggingface.co/spaces/McGill-NLP/agent-reward-bench-leaderboard", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/McGill-NLP/agent-reward-bench/raw/main/README.md"], "reason": "The dataset card supplies custom Terms of Use and preserves third-party conditions. The repository has no root code license, and no dataset license is inherited from its benchmarks."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "fb753e41c3eb19be172e9e7fccbc6bcadf44288507f82b190ca99e64144294b6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/McGill-NLP/agent-reward-bench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "7dfac983a6e443d5f66e1fc0fc3706ebea4c93005f03457cc0a2b13a2bd89601", "supports": ["identity"], "url": "https://api.github.com/repos/McGill-NLP/agent-reward-bench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "79a1639cd4738719e51fc7f32313d63c0483914517cee96175459f63513e8c07", "supports": ["release_status"], "url": "https://api.github.com/repos/McGill-NLP/agent-reward-bench/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:18Z", "role": "primary", "sha256": "20c910fad3ad45a0c8c39b182c53632a821d3b6b546278bea6b7a78fbcb3f6cb", "supports": ["release_status"], "url": "https://huggingface.co/api/datasets/McGill-NLP/agent-reward-bench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:20Z", "role": "primary", "sha256": "e16adf817e3950441d70c7bb1bb3f44bd8dc65344938247f23fea4aaff9f654e", "supports": ["licenses.data", "summary"], "url": "https://huggingface.co/datasets/McGill-NLP/agent-reward-bench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "github:mcgill-nlp/agent-reward-bench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card Terms of Use, including third-party restrictions; no standard SPDX or unrestricted permission is asserted.", "declared": "Custom AgentRewardBench Terms of Use", "retrieved_at": "2026-10-17T04:03:20Z", "url": "https://huggingface.co/datasets/McGill-NLP/agent-reward-bench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/McGill-NLP/agent-reward-bench"}], "name": "AgentRewardBench", "notes": "The public repository contains judges, evaluation utilities and annotation files; the linked Hub hosts trajectory/screenshot manifests. No root code license is declared. Dataset Terms of Use are custom and retain third-party restrictions rather than declaring an SPDX license. The paper-reported 1,302 trajectories describe its reviewed study population, not a count recomputed from current Hub files.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/McGill-NLP/agent-reward-bench/HEAD/README.md", "target_id": "arxiv:2504.08942", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A benchmark and evaluation toolkit for comparing automatic web-agent judges against expert judgments of task success, side effects and repetition.", "topics": ["evaluation", "reward-models", "web-navigation"], "updated_at": null, "url": "https://github.com/McGill-NLP/agent-reward-bench"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["web interactions"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending", "population": "safe split", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2503.04957", "unit": "tasks", "value": 250}, {"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending", "population": "harmful split", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2503.04957", "unit": "tasks", "value": 250}], "evaluation": "Task outcomes and the Agent Risk Assessment categories evaluate agent behavior on benign and harmful requests.", "evidence_urls": ["https://raw.githubusercontent.com/McGill-NLP/safearena/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"], "focus": ["web-agent safety", "task completion"], "notes": "Use SafeArena-specific containers rather than ordinary WebArena instances. The documented task splits require access approval; the profile does not imply unrestricted dataset access.", "observations": [], "setting": "interactive"}, "leaderboard_url": "https://huggingface.co/spaces/McGill-NLP/safearena-leaderboard"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T09:36:33Z", "role": "primary", "sha256": "ae50121aa60ac26b32d8d6ad6f578f9091a81a3cb9ca28b98050bddee16c929a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/McGill-NLP/safearena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "github:mcgill-nlp/safearena", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T09:36:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/McGill-NLP/safearena"}], "name": "SafeArena", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/McGill-NLP/safearena/HEAD/README.md", "target_id": "arxiv:2503.04957", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Web-agent benchmark pairing benign and harmful tasks to study deliberate misuse risks.", "topics": ["safety-security", "evaluation"], "updated_at": null, "url": "https://github.com/McGill-NLP/safearena"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["predicted browser actions"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50", "population": "original paper benchmark", "qualifier": "approximately", "unit": "interactions", "value": 100000}, {"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50", "population": "original paper benchmark", "qualifier": "approximately", "unit": "expert demonstrations", "value": 2300}, {"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50", "population": "original paper benchmark", "qualifier": "more-than", "unit": "websites", "value": 150}], "evaluation": "Offline evaluation compares action predictions against recorded human demonstrations.", "evidence_urls": ["https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50", "https://huggingface.co/datasets/McGill-NLP/WebLINX/raw/main/README.md"], "focus": ["conversational web navigation", "next-action prediction"], "notes": "The figures are rounded descriptions from the original paper. WebLINX 1.1 / BrowserGym removes some demonstrations and changes evaluated steps, including tab actions; do not mix those results or counts with the original release. This profile does not treat replayed demonstration steps as live task completion.", "observations": ["dialogue history", "HTML", "screenshots", "action history"], "setting": "offline"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "15c3f1656afd861894b180bca7d9214b4cf78a6ad9cd0601c80f5d67d2839528", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "15c3f1656afd861894b180bca7d9214b4cf78a6ad9cd0601c80f5d67d2839528", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:23Z", "role": "primary", "sha256": "9bd7512e257f5914d66b091530d7eb711a6f90bc21dc89cbba53015c7136c527", "supports": ["details.benchmark_protocol"], "url": "https://huggingface.co/datasets/McGill-NLP/WebLINX/raw/main/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "cc0156153933fb96f6f3bf8d736c15e24d78d0838945c2ed456e3ada96cf67c4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/LICENSE"}], "id": "github:mcgill-nlp/weblinx", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/mcgill-nlp/weblinx"}], "name": "WebLINX", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/README.md", "target_id": "arxiv:2402.05930", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Demonstrations and evaluation for conversational web navigation.", "topics": ["evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/mcgill-nlp/weblinx"}
{"aliases": [], "authors": [], "details": {"availability": "Public fine-tuning/evaluation tooling, examples and results viewer.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/LICENSE", "https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/README.md"], "reason": "README explicitly scopes MIT to code unless a file header states otherwise; models/data/images retain original licenses."}}, "repository_revision": "1b7b41327b2222312e07818ce239c04307beaf9b"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "121546c7a73f4d7604ae5340bd78856e8170dc8aaa8e57f7e62efcfb85d8d4c9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "6605b7b09e8be77cce9d5bfd1a6199c2256cbfc1f067b93b2b952c1d19e74f10", "supports": ["identity"], "url": "https://api.github.com/repos/McGill-NLP/webllama"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "915620d7a155f5c5fe099cdb0461d9614e1894b38559f0d2f9a6a138afe06212", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/McGill-NLP/webllama/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "e69268828e67ffd1d3393a2d3530f97df6deaa0b961039fc971abea28cec8715", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/LICENSE"}], "id": "github:mcgill-nlp/webllama", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:55Z", "licenses": {"code": {"basis": "README explicitly scopes MIT to code unless a file header states otherwise; models/data/images retain original licenses.", "declared": "MIT", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/McGill-NLP/webllama"}], "name": "WebLlama", "notes": "The manifest contains modeling/training code, BrowserGym examples, documentation and a results viewer. The README distinguishes currently available components from planned additional benchmarks and deployment integrations. It requests citation of WebLINX as the foundation for data, training and evaluation; this is not a claim that the WebLINX paper first introduced every later WebLlama component. MIT applies to repository code, subject to file-header exceptions; weights and datasets have their own terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/README.md", "target_id": "arxiv:2402.05930", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training, evaluation and integration tooling for dialogue-guided web agents, including Llama fine-tuning workflows and BrowserGym examples built around WebLINX.", "topics": ["web-navigation", "imitation-learning", "evaluation", "tool-integration"], "updated_at": null, "url": "https://github.com/McGill-NLP/webllama"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "b00a072dfde61f774db6acf94fcde9bd32337b1b4f16a178998378d842bbbe4d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/meituan/EvoCUA/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "a8d1d88a0f9cce3cdbff73b03a80dc5e5d69e3ba2443946c268815a550c9e0ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/meituan/EvoCUA/HEAD/LICENSE"}], "id": "github:meituan/evocua", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/meituan/EvoCUA/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/meituan/EvoCUA"}], "name": "EvoCUA", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/meituan/EvoCUA/HEAD/README.md", "target_id": "arxiv:2601.15876", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Computer-use model and training resources from the EvoCUA project.", "topics": ["reinforcement-learning", "planning"], "updated_at": null, "url": "https://github.com/meituan/EvoCUA"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "ce11bdfe8354eb9c3b8a985dd5ddd21f88f94fb5fb9d00c9b01a76226f0b0a36", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/fara/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "98a96c38494df4963951e32b81ec4effebe4c8a812ab8d9e7d185d59e86fe2fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/fara/HEAD/LICENSE"}], "id": "github:microsoft/fara", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/microsoft/fara/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/fara"}], "name": "Fara", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/fara/HEAD/README.md", "target_id": "arxiv:2511.19663", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/microsoft/fara/HEAD/README.md", "target_id": "arxiv:2606.20785", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Microsoft browser-use agent models that act from screenshots by predicting click coordinates without accessibility trees; Fara1.5 comes in 4B, 9B and 27B, succeeding Fara-7B.", "topics": ["planning", "web-navigation"], "updated_at": null, "url": "https://github.com/microsoft/fara"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dafdebf2fa1e0a5db871addd103e50cb55625f0108c757708029e885920ee4e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "98a96c38494df4963951e32b81ec4effebe4c8a812ab8d9e7d185d59e86fe2fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/LICENSE"}], "id": "github:microsoft/gui-actor", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/GUI-Actor"}], "name": "GUI-Actor", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/README.md", "target_id": "arxiv:2506.03143", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI grounding model with attention-based action localization.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/microsoft/GUI-Actor"}
{"aliases": [], "authors": [], "details": {"availability": "Public training and evaluation source; trained checkpoints are generated by the documented workflow.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/LICENSE"], "reason": "Repository LICENSE grants MIT for code; linked training imagery and pretrained model licenses remain separate."}}, "repository_revision": "6d2381ef549d980452a828471221a2ce8023bf41"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:14Z", "role": "primary", "sha256": "d21a1ae4359296c0f3034ac655888dce43448d10afe0b1909d6b6520678c2526", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:13Z", "role": "primary", "sha256": "791c6e56e62ce5e803a4ba313557bac0a39dfeba27499a3865ad444a860b426a", "supports": ["identity"], "url": "https://api.github.com/repos/microsoft/GUI-Agent-RL"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:06Z", "role": "primary", "sha256": "5b799a5bdae78ff50652c0b464b6cad12d9dda265d4752e5d5f9eab65e9d5c6f", "supports": ["release_status"], "url": "https://api.github.com/repos/microsoft/GUI-Agent-RL/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:06Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/LICENSE"}], "id": "github:microsoft/gui-agent-rl", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T09:12:14Z", "licenses": {"code": {"basis": "Repository LICENSE grants MIT for code; linked training imagery and pretrained model licenses remain separate.", "declared": "MIT", "retrieved_at": "2026-10-17T09:14:06Z", "url": "https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/GUI-Agent-RL"}], "name": "VEM GUI-Agent-RL", "notes": "The public tree includes critic-training configurations, policy training and offline/online evaluation scripts. The README obtains base models and AITW imagery separately, then instructs users to train critic and policy checkpoints locally; this entry does not claim that VEM-trained weights are released. MIT applies to repository code only.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/README.md", "target_id": "arxiv:2502.18906", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training code that learns a value environment model from offline Android interaction data and uses its frozen value estimates to guide GUI policy optimization.", "topics": ["reinforcement-learning", "reward-models", "mobile-navigation"], "updated_at": null, "url": "https://github.com/microsoft/GUI-Agent-RL"}
{"aliases": ["Magentic UI", "Magentic-UI", "MagenticLite"], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "240e0c92b86c02d3790c0784d8c72193ffb4fd5f9c2b0ec618ef8121fe216b8b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/magentic-ui/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "98a96c38494df4963951e32b81ec4effebe4c8a812ab8d9e7d185d59e86fe2fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/magentic-ui/HEAD/LICENSE"}], "id": "github:microsoft/magentic-ui", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/microsoft/magentic-ui/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/magentic-ui"}], "name": "MagenticLite / Magentic UI", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Microsoft agent application combining orchestration with browser interaction; successor to Magentic UI.", "topics": ["human-in-the-loop"], "updated_at": null, "url": "https://github.com/microsoft/magentic-ui"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:18:21Z", "role": "primary", "sha256": "8aaf8feff6e1d5a533c01afd8afc0cfc0700f0c1c73ba47949b56c11fb8ab5be", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Magma/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T17:18:56Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/Magma/HEAD/LICENSE"}], "id": "github:microsoft/magma", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T17:18:21Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T17:18:56Z", "url": "https://raw.githubusercontent.com/microsoft/Magma/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/Magma"}], "name": "Magma", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Microsoft foundation model for multimodal agents, pretrained with Set-of-Mark and Trace-of-Mark objectives on UI, robotics and video data and fine-tunable for web and mobile UI navigation.", "topics": ["grounding", "planning", "mobile-navigation"], "updated_at": null, "url": "https://github.com/microsoft/Magma"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "f4ebaafcc0ab0b9e9cc55cb587678e5c83a5bbd64469fb2cc52bc63843c5cc6d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/OmniParser/HEAD/README.md"}], "id": "github:microsoft/omniparser", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/OmniParser"}], "name": "OmniParser", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/OmniParser/HEAD/README.md", "target_id": "arxiv:2408.00203", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screenshot parser that detects interactable regions and captions icons so vision-language models can ground actions, released as V1.5 and V2 weights alongside OmniTool for driving a Windows 11 VM.", "topics": ["screen-understanding", "grounding"], "updated_at": null, "url": "https://github.com/microsoft/OmniParser"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a0609394a736dc495d077cd57a2945b018123e5b3f9514d7d5bafb9e690aeae2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/LICENSE"}], "id": "github:microsoft/phi-ground", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/Phi-Ground"}], "name": "Phi-Ground", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md", "target_id": "arxiv:2507.23779", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Microsoft GUI grounding models from an empirical study of grounding training, from data collection to model training, released as Phi-Ground-4B-7C with evaluation code.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/microsoft/Phi-Ground"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "0a6a9dddcbf1295b61e83462ee7aa130289a4d4e60bac1a0deccd661f8fc8146", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/playwright/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "45873d00a0dd243596deb4aa23b2493b3d1f0671921bf2538ea431d7380220eb", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/playwright/HEAD/LICENSE"}], "id": "github:microsoft/playwright", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/microsoft/playwright/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/playwright"}], "name": "Playwright", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Browser automation library used as an execution layer by agents.", "topics": ["automation"], "updated_at": null, "url": "https://github.com/microsoft/playwright"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "b33aa71a13d655e96e6b736d0c9de1e5ddea52f205916ca20b876387ff830e87", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/playwright-mcp/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "9a7110fc2d2f964038e5dc49128f908f29f47a574c961cba16085914e879cbda", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/playwright-mcp/HEAD/LICENSE"}], "id": "github:microsoft/playwright-mcp", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/microsoft/playwright-mcp/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/playwright-mcp"}], "name": "Playwright MCP", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "MCP server that lets LLM clients drive Playwright browsers through structured accessibility snapshots instead of screenshots, with opt-in vision, PDF and devtools capabilities.", "topics": ["automation", "tool-integration"], "updated_at": null, "url": "https://github.com/microsoft/playwright-mcp"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["pixel (x,y) cursor coordinate"], "counts": [], "evaluation": "The inspected evaluator scales target boxes from a 1000×1000 reference grid into image pixels and tests pixel predictions, allowing ±10 pixels around the center along any target dimension narrower than 1 pixel after scaling. It reports cumulative hit accuracy by turn (earlier hits remain correct), miss distances, any-turn hit rate, and a correction rate whose denominator is all successfully processed samples. The default budget is five turns; retain prompt/feedback variant, turn budget, and processed/skipped coverage.", "evidence_urls": ["https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/README.md", "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/eval/test_allturns_api.py", "https://api.github.com/repos/microsoft/precision-cua-bench/git/trees/main?recursive=1"], "focus": ["pixel-precise editor cursor placement", "multi-turn visual self-correction"], "notes": "The release has four editor/theme JSONL partitions (Cursor/VS Code × dark/light) and screenshot files. Partitions are not train/test splits or independent task totals. README’s initial field table calls bbox pixel coordinates, but the evaluation-format section and actual ratio_h/ratio_w=dimension/1000 code use normalized target boxes; profile follows the inspected evaluator and retains the inconsistency. Source/images were not executed or downloaded as a dataset. MIT is recorded for repository code only.", "observations": ["code-editor screenshot", "cursor-location instruction", "red-cross feedback marking previous incorrect prediction"], "setting": "offline"}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:52Z", "role": "primary", "sha256": "23d0dc8ea0a10c080cf8ae32121eade2785ae923a6b0d853d166c94c195e789b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://github.com/microsoft/precision-cua-bench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "84ad4fb798abb776dce813cf4dab296d6e03248d8ae29388e1f06cccd27d1241", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:08Z", "role": "primary", "sha256": "1fda7c3980fe7a7c373478eda6ecfabff227e92c2b1f19c6fe903b408f583797", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/eval/test_allturns_api.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:06Z", "role": "primary", "sha256": "ede84e3dda47f80fada4234caa2423db08e21ff7667a37599c03aae657b2f84b", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/microsoft/precision-cua-bench/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/LICENSE"}], "id": "github:microsoft/precision-cua-bench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:52Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/precision-cua-bench"}], "name": "Precision CUA Bench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50", "target_id": "arxiv:2604.13019", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Evaluation resources for pixel-precise cursor grounding in code editors.", "topics": ["evaluation", "grounding"], "updated_at": null, "url": "https://github.com/microsoft/precision-cua-bench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "b3fa8b7e24a01da023a5f21023c4853a90388dd7744055fbb5a2888938c68b0c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/LICENSE"}], "id": "github:microsoft/som", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/SoM"}], "name": "Set of Mark", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/README.md", "target_id": "arxiv:2310.11441", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "Visual-prompting toolbox that labels image regions for multimodal reference and grounding, including GUI applications.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/microsoft/SoM"}
{"aliases": [], "authors": [], "details": {}, "environments": ["windows", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "40d9a4010ab122ab50242a0cd7a5320e2fd16b3e53494d988eecdbd94fd8e6e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/UFO/HEAD/README.md"}], "id": "github:microsoft/ufo", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/UFO"}], "name": "UFO", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/UFO/HEAD/README.md", "target_id": "arxiv:2402.07939", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Windows agent framework for application and desktop automation.", "topics": ["planning", "cross-app"], "updated_at": null, "url": "https://github.com/microsoft/UFO"}
{"aliases": ["Webwright", "CUAWright"], "authors": [], "details": {}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "7c0f87f684e5ed4d77b175b63a5dcaf0a268644c3662c723713de4cf0ac1de40", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Webwright/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/Webwright/HEAD/LICENSE"}], "id": "github:microsoft/webwright", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/microsoft/Webwright/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/Webwright"}], "name": "CUAWright (formerly Webwright)", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/Webwright/HEAD/README.md", "target_id": "arxiv:2610.04116", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Agent framework for browser and desktop tasks using automation code.", "topics": ["tool-integration"], "updated_at": null, "url": "https://github.com/microsoft/Webwright"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["agent-defined Windows interactions"], "counts": [{"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_all.json", "population": "evaluation_examples_windows/test_all.json", "qualifier": "exact", "unit": "task definitions", "value": 154}], "evaluation": "Task configuration selects an evaluator and expected application or system outcome; evaluators return a reward for the resulting state.", "evidence_urls": ["https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/README.md", "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/docs/Develop-Tasks.md", "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_all.json"], "focus": ["Windows task completion", "desktop workflows"], "notes": "Counted unique identifiers in the inspected manifest. Normal and hard modes differ in how much task setup the agent must perform. Record the difficulty and observation/backend configuration with each result; no VM was launched.", "observations": ["screenshots", "accessibility information", "screen-element detections"], "setting": "interactive"}}, "environments": ["windows", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "e481abd422a4cc8b819ad84b32684e4af641e18a7c0d277dd608cfd5d3c5e6ad", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "e481abd422a4cc8b819ad84b32684e4af641e18a7c0d277dd608cfd5d3c5e6ad", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:06:30Z", "role": "primary", "sha256": "c565dc02ab0049bf9710b87e2760dda8d7c92268ea78f0354952695ab59c0de3", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/docs/Develop-Tasks.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:06:30Z", "role": "primary", "sha256": "902fb24cf0a9e146bf6c29288e74a7254fe40294e49dd47363e0b15d65608fef", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_all.json"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "9906940f61b1f0b533fa7d99baf55178b2808fbe113ea51dfbfad8572ccd5f2b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/LICENSE"}], "id": "github:microsoft/windowsagentarena", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/WindowsAgentArena"}], "name": "Windows Agent Arena", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/README.md", "target_id": "arxiv:2409.08264", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Windows 11 benchmark for multimodal desktop agents in a reproducible VM that runs locally or in parallel on Azure ML, with a harder mode where agents must set tasks up themselves.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/microsoft/WindowsAgentArena"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["Selenium-backed browser interactions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md", "population": "WebVoyager task collection", "qualifier": "exact", "unit": "task queries", "value": 643}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md", "population": "WebVoyager task collection", "qualifier": "exact", "unit": "websites", "value": 15}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md", "population": "separate GAIA validation-derived web subset", "qualifier": "exact", "unit": "tasks", "value": 90}], "evaluation": "The repository supplies an automated multimodal-model evaluation protocol.", "evidence_urls": ["https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md"], "focus": ["live-web task completion"], "notes": "The GAIA-derived subset is separate from the 643 WebVoyager queries. Live sites and time-sensitive travel tasks can change; record the task snapshot and evaluation date.", "observations": ["webpage screenshots", "textual task instructions"], "setting": "interactive"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "b5523941b2773634fdcf359b5b89bf0aca589f33ad9db30692c8b1912ed5dda5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "43070e2d4e532684de521b885f385d0841030efa2b1a20bafb76133a5e1379c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/LICENSE"}], "id": "github:minorjerry/webvoyager", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MinorJerry/WebVoyager"}], "name": "WebVoyager", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md", "target_id": "arxiv:2401.13919", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Multimodal web navigation agent and evaluation tasks on real websites.", "topics": ["evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/MinorJerry/WebVoyager"}
{"aliases": [], "authors": [], "details": {"availability": "Published source; official release metadata inspected", "release_tag": "1.0.11"}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "bc24fa665d1a49fa806d58ec569d93e134058e0cb1a7bc9324339ba0e35ff107", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/mobile-next/mobile-mcp/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "80e8ad254f71f5c80b91a6b80edd09f96116ef4967ae4e6832e5960357f7b99a", "supports": ["identity"], "url": "https://api.github.com/repos/mobile-next/mobile-mcp"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "49684448428c28afd4bf191f695c3d45fbe46910fa93951d2a02f6b8afcf9f16", "supports": ["release_status"], "url": "https://api.github.com/repos/mobile-next/mobile-mcp/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:57Z", "role": "primary", "sha256": "43a5c02407afe86b6f1fedb4299f28420778ce21825cd87516deafe05507a7bc", "supports": ["release_status"], "url": "https://api.github.com/repos/mobile-next/mobile-mcp/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:58Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/mobile-next/mobile-mcp/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:00:58Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/mobile-next/mobile-mcp/HEAD/LICENSE"}], "id": "github:mobile-next/mobile-mcp", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T03:59:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T04:00:58Z", "url": "https://raw.githubusercontent.com/mobile-next/mobile-mcp/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/mobile-next/mobile-mcp"}], "name": "Mobile Next MCP", "notes": "Source and release 1.0.11 are public. The README specifies iOS and Android device setup and supports local devices as well as a separate cloud offering. Apache-2.0 is declared for this code; no cloud-service pricing or permissions are inferred.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "An MCP server exposing mobile accessibility snapshots, screenshots and device interactions across real devices, simulators and emulators.", "topics": ["mobile-navigation", "tool-integration", "automation"], "updated_at": null, "url": "https://github.com/mobile-next/mobile-mcp"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "e1d75e8a196f19af6050c53694383f069561244c810ca259c5f8075ee91f72b1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MobileAgentBench/mobile-agent-bench/HEAD/README.md"}], "id": "github:mobileagentbench/mobile-agent-bench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MobileAgentBench/mobile-agent-bench"}], "name": "MobileAgentBench", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Automated benchmark for mobile LLM agents whose default tasks use ten Simple Mobile Tools Android apps, installed as a Python library that detects task completion.", "topics": ["evaluation", "mobile-navigation"], "updated_at": null, "url": "https://github.com/MobileAgentBench/mobile-agent-bench"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "ef5e2998ac2defa86e0dc4c13e7e00d595e1e7cf", "repository_url": "https://github.com/MobileLLM/AgentProg"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:13Z", "role": "primary", "sha256": "becc4961d93f0170fab29120afa4b09b2999f2cd0e74b13ccb7f28cb9623761b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:01Z", "role": "primary", "sha256": "d14eecf7c041933cefecc03b62bbea45c1617a183fd8c4ec1334e22a644a7f67", "supports": ["identity"], "url": "https://api.github.com/repos/MobileLLM/AgentProg"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:22Z", "role": "primary", "sha256": "dd1cf9cd0aed31af63d7046ffa98d1be2c36258706951173ad8c28a64fcf9f7a", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/MobileLLM/AgentProg/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "751123fccb2fd04b5c08b8c712a68a8bc69f56391593711efbfa533eaebbdb10", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/LICENSE"}], "id": "github:mobilellm/agentprog", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:13Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MobileLLM/AgentProg"}], "name": "AgentProg", "notes": "Public tree contains the agentprog package, CLI, program/context-management code and AndroidWorld evaluation integration. README announces AW-Extend, but this record is the framework rather than a separately audited benchmark release. External model services and Android device/emulator setup are required. Apache-2.0 covers code, not external model weights, apps or VM images.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/README.md", "target_id": "arxiv:2512.10371", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Android GUI-agent framework that keeps long-running task context in structured programs, variables and control flow, with a CLI and AndroidWorld integration.", "topics": ["long-horizon", "memory", "planning", "state-tracking"], "updated_at": null, "url": "https://github.com/MobileLLM/AgentProg"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "4a79eb2e14b6da67ded536c127d41f865d5b43620d60d0d950d291c8956564fc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid/HEAD/README.md"}], "id": "github:mobilellm/autodroid", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MobileLLM/AutoDroid"}], "name": "AutoDroid", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid/HEAD/README.md", "target_id": "arxiv:2308.15272", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "LLM-powered Android task automation built on DroidBot, released with DroidTask, a dataset of per-app screenshots, view hierarchies, exploration records and ground-truth task files.", "topics": ["mobile-navigation"], "updated_at": null, "url": "https://github.com/MobileLLM/AutoDroid"}
{"aliases": [], "authors": [], "details": {"availability": "Public source and minimal-experiment documentation are released; default README path is absent.", "repository_revision": "b1a02a52a6c94b96f63016ac12d3a4ab6a410b8e"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "74508100843f8db6db6bb151e77ac1cc604a6b7ebd16d3225266d73434465136", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/README_minimal.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:10Z", "role": "primary", "sha256": "f349963f2ea3f7b81f1c04ee875654382929b1f93d99d0168d6e086155f41d72", "supports": ["identity"], "url": "https://api.github.com/repos/MobileLLM/AutoDroid-V2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:03Z", "role": "primary", "sha256": "70354b856b75bc4a2aca0a7274321637039ec5d083523ca2f677cd11d475b386", "supports": ["release_status", "licenses.code"], "url": "https://api.github.com/repos/MobileLLM/AutoDroid-V2/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "8ba37983a2c2da53d96a52696c07e4bdbea4e987a1a07b557935d614c3be46d2", "supports": ["summary", "details", "relations"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_3_training/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "cfed016e038d09ce7c9e829273fffa4a6c618bf8e9043312e4690fbe1f40d247", "supports": ["environments", "details"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_0_installation/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "f5534771f68770ed2e45daf94fbf80304ba2904387571c34c5d80a152bef0a00", "supports": ["summary"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_1_doc_generation/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:59Z", "role": "primary", "sha256": "0564f566fde97360496bea4a6eb023c4c84e03cdfdf5201c5b27235c36e32b18", "supports": ["summary", "details"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_4_accuracy_validation/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "github:mobilellm/autodroid-v2", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T09:14:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MobileLLM/AutoDroid-V2"}], "name": "AutoDroid V2", "notes": "The repository uses README_minimal.md rather than README.md and includes a documented App Launcher experiment, training setup and DroidTask/LlamaTouch evaluation scripts. Its supplied emulator setup targets x86_64 Linux and Android API 31. No repository-level code license was found; code, APK, data and weight terms are not inferred from one another. Dependencies, external evaluation files and model service are still required; no software was run.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2412.18116", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An Android automation implementation that builds app documentation from exploration traces and evaluates language-model-generated execution code in emulator workflows.", "topics": ["mobile-navigation", "tool-integration", "data-generation"], "updated_at": null, "url": "https://github.com/MobileLLM/AutoDroid-V2"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:34Z", "role": "primary", "sha256": "3095f7262b7d4ff916c6f58c55e6fc9902c08f77d09bf3951199945501148bb3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/LICENSE"}], "id": "github:murtyshikhar/nnetnav", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:42:34Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MurtyShikhar/NNetnav"}], "name": "NNetNav", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md", "target_id": "arxiv:2410.02907", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Collects and labels browser-exploration trajectories for agent training, with code for data processing, fine-tuning, and web evaluation.", "topics": ["data-generation", "web-navigation"], "updated_at": null, "url": "https://github.com/MurtyShikhar/NNetnav"}
{"aliases": [], "authors": [], "details": {"availability": "Published source; official release metadata inspected", "release_tag": "v0.3.0"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "4d8a0b0d6286c4bab116c155f087948062166e7e169d6bbaf504fedb4f32dc67", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/nanobrowser/nanobrowser/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "6b48386f32a971a724906de8b6d5ea53036f3813411ddc12931be3010c034400", "supports": ["identity"], "url": "https://api.github.com/repos/nanobrowser/nanobrowser"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "afd0b71b01d3d89033d8f32abaf9871f38e8fb0c50442ed3a8cbb3079d4a4fe0", "supports": ["release_status"], "url": "https://api.github.com/repos/nanobrowser/nanobrowser/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:00Z", "role": "primary", "sha256": "c7b28db1b1c8e78d100da066b08db713bc144baff99941b6c94489ef61baee8f", "supports": ["release_status"], "url": "https://api.github.com/repos/nanobrowser/nanobrowser/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/nanobrowser/nanobrowser/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:01:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/nanobrowser/nanobrowser/HEAD/LICENSE"}], "id": "github:nanobrowser/nanobrowser", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T04:01:00Z", "url": "https://raw.githubusercontent.com/nanobrowser/nanobrowser/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/nanobrowser/nanobrowser"}], "name": "Nanobrowser", "notes": "The official repository publishes extension source and a v0.3.0 ZIP release. It names Chrome and Edge as supported browsers. Apache-2.0 is declared for the code. Ask/Auto behavior and provider choices are documented, but no security guarantee, model performance or service-price claim is inferred.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A browser-extension agent that reads pages, interacts with elements and moves between tabs from a side-panel task interface.", "topics": ["web-navigation", "automation", "human-in-the-loop"], "updated_at": null, "url": "https://github.com/nanobrowser/nanobrowser"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md", "population": "reported MultiUI training collection", "qualifier": "approximately", "unit": "samples", "value": 7300000}], "evidence_urls": ["https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md"], "formats": [], "modalities": ["webpage UI content", "text", "accessibility-derived structure"], "notes": "The README provides a rounded sample count. It does not establish a unique-screenshot inventory or a normalized split breakdown. Synthetic examples and publicly sourced web content have provenance and quality limits; data terms must be checked separately from repository code.", "splits": ["training collection"], "supervision": ["UI understanding", "text-rich visual tasks"]}, "modalities": ["screenshot", "instruction"], "reported_scale": {"comparison": "approximately", "unit": "samples", "value": 7300000}, "sample_count": null, "sample_unit": "instructions"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "fe0c03fe73529a29788d559103c4aed48dca6ed9ba3bfd5f151f902fa8858b38", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "fe0c03fe73529a29788d559103c4aed48dca6ed9ba3bfd5f151f902fa8858b38", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md"}], "id": "github:neulab/multiui", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/neulab/multiui"}], "name": "MultiUI", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "7.3 million training samples built from webpage UIs and structured with enhanced accessibility trees, for text-rich visual understanding, released with UIX-Qwen2 models.", "topics": ["screen-understanding", "data-generation"], "updated_at": null, "url": "https://github.com/neulab/multiui"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:20Z", "role": "primary", "sha256": "2fb7c0d70ccfa93605d5250a8f6d5182324f0c53d70e1c35ba18cb00f4ada548", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/night-chen/WorkForceAgent-R1/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:19Z", "role": "primary", "sha256": "0ad3ef1d900200e3ecab86cd44f79038000e8748fb285cdd5f32882455d2e36e", "supports": ["identity"], "url": "https://api.github.com/repos/night-chen/WorkForceAgent-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:39Z", "role": "primary", "sha256": "521d065188a6354416c1aafe88a0eb34486f2c2ec011c1cdf8ed008760dd9c55", "supports": ["release_status"], "url": "https://api.github.com/repos/night-chen/WorkForceAgent-R1/git/trees/main?recursive=1"}], "id": "github:night-chen/workforceagent-r1", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:03:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/night-chen/WorkForceAgent-R1"}], "name": "WorkForceAgent-R1", "notes": "The release contains WorkArena preprocessing and training source. The inspected root has no explicit project license; dependencies retain their own terms. Although the README discusses trained 7B/14B models, it does not provide a verified checkpoint release, so this is cataloged only as a framework.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/night-chen/WorkForceAgent-R1/HEAD/README.md", "target_id": "arxiv:2505.22942", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Web-task reward functions, WorkArena data processing and GRPO training configurations for reasoning-oriented language agents.", "topics": ["reinforcement-learning", "web-navigation", "planning"], "updated_at": null, "url": "https://github.com/night-chen/WorkForceAgent-R1"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "2248b2048b9166ce0348b24e81b2587fdf373a958fd86ab870ba85825425b02f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/niuzaisheng/ScreenAgent/HEAD/README.md"}], "id": "github:niuzaisheng/screenagent", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/niuzaisheng/ScreenAgent"}], "name": "ScreenAgent", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/niuzaisheng/ScreenAgent/HEAD/README.md", "target_id": "arxiv:2402.07945", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Environment and agent in which a vision-language model controls real desktops over VNC with mouse and keyboard in plan, act and reflect stages, plus a hand-annotated dataset of daily computer tasks.", "topics": ["planning"], "updated_at": null, "url": "https://github.com/niuzaisheng/ScreenAgent"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "c4112415bec58fc77f9a1e8fa98f8d602c5d382747579dfda9f866900c55995c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "43070e2d4e532684de521b885f385d0841030efa2b1a20bafb76133a5e1379c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/LICENSE"}], "id": "github:njucckevin/seeclick", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/njucckevin/SeeClick"}], "name": "SeeClick", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md", "target_id": "arxiv:2401.10935", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI grounding model trained to locate visual interface elements.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/njucckevin/SeeClick"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "5ecb73264cce10b75ec04dea59aaff77ffde552259fba6932507779a35681c5b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/nottelabs/notte/HEAD/README.md"}], "id": "github:nottelabs/notte", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/nottelabs/notte"}], "name": "Notte", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Web agent framework and cloud SDKs for managed Chrome sessions, adding a CLI, an MCP server for coding agents, serverless browser functions and an API that builds automations from English tasks.", "topics": ["web-navigation"], "updated_at": null, "url": "https://github.com/nottelabs/notte"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "41a6435d6c27a647b4d2170cba476fa6a941b12f8a4ffe41674a508df9e0ea1d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/nut-tree/nut.js/HEAD/README.md"}], "id": "github:nut-tree/nut.js", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/nut-tree/nut.js"}], "name": "Nut.js", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Desktop automation library for keyboard, mouse, and screen interaction.", "topics": ["automation"], "updated_at": null, "url": "https://github.com/nut-tree/nut.js"}
{"aliases": ["Oculix", "SikuliX continuation"], "authors": [], "details": {"availability": "Public source and official stable release assets.", "license_components_note": "Project source declares MIT; the retained third-party dependency notices describe additional component terms.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/LICENSE"], "reason": "Publisher project-source LICENSE. Third-party dependency notices are separately retained and no blanket MIT assertion is made for the compiled distribution."}}, "release_tag": "v4.0.0", "repository_revision": "02ea8844483a83a2963db8016cd7ad421e15bc91"}, "environments": ["desktop", "windows", "macos", "linux", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:14Z", "role": "primary", "sha256": "41d32f40bfb960f38b77ebb62a73ae20454be6c52829c9cbf0b8fedd27de2297", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:14Z", "role": "primary", "sha256": "151ed1aaf544b9afcfd2d8f13d0fa682d8cf8f8a68496af20c3597b082070801", "supports": ["identity"], "url": "https://api.github.com/repos/oculix-org/Oculix"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:48Z", "role": "primary", "sha256": "cba4541e5d89210b570e42cc69c1529ca36e698622c69b4bb0eeffe56339a759", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/oculix-org/Oculix/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:08Z", "role": "primary", "sha256": "1022eb9591a2f01833e88c2bae2869faeac9b92e535901d2c79e93a714278777", "supports": ["release_status", "details.release_tag", "environments", "notes"], "url": "https://api.github.com/repos/oculix-org/Oculix/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:14Z", "role": "primary", "sha256": "8994ff2000b0871aca819ecd7233430bb82cea3f0134a555cebbfa9e02e0f4da", "supports": ["identity", "relations", "notes"], "url": "https://raw.githubusercontent.com/oculix-org/SikuliX1/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:13Z", "role": "primary", "sha256": "102b86e9019fe5e936486b3213d2bbd9e80d6c7ad393284b1a4298b0be555ce7", "supports": ["identity"], "url": "https://api.github.com/repos/RaiMan/SikuliX1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:03Z", "role": "primary", "sha256": "cea4b6292d59625e14968ec666faa7ae6e237b510d846a96875bd6a282677f3a", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:27:50Z", "role": "primary", "sha256": "f0e0d8aa91c1c60bac2250688248c85503036aeac37ec02b8a75fe9d50b6ae4c", "supports": ["licenses.code", "notes"], "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/API/src/main/resources/Settings/THIRD-PARTY.txt"}], "id": "github:oculix-org/oculix", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T11:18:14Z", "licenses": {"code": {"basis": "Publisher project-source LICENSE. Third-party dependency notices are separately retained and no blanket MIT assertion is made for the compiled distribution.", "declared": "MIT", "retrieved_at": "2026-10-17T11:23:03Z", "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/oculix-org/Oculix"}], "name": "OculiX (SikuliX continuation)", "notes": "The official stable-release API lists v4.0.0 with platform-specific API/IDE JARs and an MCP server JAR; no binary was downloaded. The default README still advertises 3.0.3, so release-specific compatibility follows the release notes rather than that older badge. The RaiMan/SikuliX1 API resolves to oculix-org/SikuliX1, whose README calls itself a historical mirror and directs current users here; it is not separately counted. MIT describes the project source. Retained dependency notices also include GPL/LGPL, Apache, BSD and MPL declarations; a bundled artifact is not asserted to be uniformly MIT.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/README.md", "target_id": "github:sikuli/sikuli", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "A maintained continuation of SikuliX that locates on-screen targets with image matching and OCR, drives desktop or Android interfaces, and exposes scripting and optional MCP interfaces.", "topics": ["automation", "grounding", "tool-integration"], "updated_at": null, "url": "https://github.com/oculix-org/Oculix"}
{"aliases": [], "authors": [], "details": {"availability": "Public browser-agent source and npm release metadata.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/omxyz/lumen/HEAD/LICENSE"], "reason": "Publisher project-source license; external model and browser-service terms are separate."}}, "package_version": "0.2.0", "repository_revision": "b1ad26a0784645ac3a97d402db99cd5d17f86334"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "7708f434142498959bb5ed5de35ac1b945f9ad6e070e436b4e73819b7db1df7f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/omxyz/lumen/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "32b02ea43d19929d82e8b8c52bb8e448553aebdd96a8754d144927cb0f34c5b6", "supports": ["identity"], "url": "https://api.github.com/repos/omxyz/lumen"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "9911953e501a245d657f3835b22c6822ad7d990b96f03e3cba93a15ebe4f70a6", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/omxyz/lumen/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "34a03b6dd8cd1960d5592a6569e2a1c1f6c1b2697da1386a58d0ca64157f059e", "supports": ["identity", "release_status", "details.package_version", "notes"], "url": "https://registry.npmjs.org/@omxyz%2Flumen/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "30dd5336a8a45b548b9051e751d45923a9864b3810a36c836df9b07447ca3c5c", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/omxyz/lumen/HEAD/LICENSE"}], "id": "github:omxyz/lumen", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T12:30:55Z", "licenses": {"code": {"basis": "Publisher project-source license; external model and browser-service terms are separate.", "declared": "MIT", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/omxyz/lumen/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/omxyz/lumen"}], "name": "Lumen browser agent", "notes": "Public TypeScript source and npm package metadata are available; the inspected package version is 0.2.0. The package declares Node ^20.19.0 or >=22.12.0 and requires a browser and model provider. Its README benchmark is explicitly preliminary and uses a small WebVoyager subset; this catalog makes no comparative performance or independent safety claim. Referenced research influences are not asserted as papers introducing Lumen.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A screenshot-driven browser agent over Chrome DevTools Protocol, with resumable sessions, context compression, action replay and configurable completion checks.", "topics": ["web-navigation", "memory", "automation", "observability"], "updated_at": null, "url": "https://github.com/omxyz/lumen"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": [], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md", "population": "reported multi-level benchmark", "qualifier": "more-than", "unit": "tasks", "value": 8000}], "evaluation": "Level-specific evaluation; Efficiency–Quality Area (EQA) combines navigation quality and efficiency.", "evidence_urls": ["https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md"], "focus": ["screen understanding", "grounding", "task automation", "task collaboration"], "notes": "The README documents four levels, but its release notes and TODOs distinguish released L1/L2 evaluation from pending L3/L4 integration. Do not treat every level as the same evaluator or a complete implemented interface.", "observations": ["screenshots"], "setting": "mixed"}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "35db9fca3ed1c1a3513924fefff4ccd224bc70cd7eb4bf8c54fcb9048851988e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md"}], "id": "github:open-compass/mmbench-gui", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/open-compass/MMBench-GUI"}], "name": "MMBench-GUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md", "target_id": "arxiv:2507.19478", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Four-level GUI-agent benchmark across six platforms, from content understanding and element grounding to task automation and collaboration, scored with an efficiency-quality area metric.", "topics": ["evaluation", "grounding"], "updated_at": null, "url": "https://github.com/open-compass/MMBench-GUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "36a6e7d56a1d474ee83e5811babe8e11cb113ee12386f9621bba64ffd957a6bc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenAdaptAI/OpenAdapt/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "7f2bc0ccbf15854bbe46f5fa0e7cb26fe42f88ec9f8c9b151f0f93c4415820f9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenAdaptAI/OpenAdapt/HEAD/LICENSE"}], "id": "github:openadaptai/openadapt", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/OpenAdaptAI/OpenAdapt/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenAdaptAI/OpenAdapt"}], "name": "OpenAdapt", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Tools for learning and automating workflows from desktop demonstrations.", "topics": ["recording", "workflow"], "updated_at": null, "url": "https://github.com/OpenAdaptAI/OpenAdapt"}
{"aliases": [], "authors": [], "details": {"availability": "Public archived client source; remote benchmark services and images are not verified usable.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/openai/universe/HEAD/LICENSE"], "reason": "Official Universe client-source LICENSE; no license or availability claim is transferred to external remote applications or container contents."}}, "maintenance_status": "archived and deprecated", "repository_revision": "cc9ce6ec241821bfb0f3b85dd455bd36e4ee7a8c"}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:17Z", "role": "primary", "sha256": "b1ca938480d4d36c76fa939c5b10cef11c24465cc1a6a8ee857cec4b939f9d61", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/openai/universe/HEAD/README.rst"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:15Z", "role": "primary", "sha256": "94a7a4d383f754b96601bc0486b90e62760b8bfc35abc607077c8eb6e7934b2d", "supports": ["identity", "details.availability", "notes"], "url": "https://api.github.com/repos/openai/universe"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:47Z", "role": "primary", "sha256": "1808d7524506d1386dad507de8a8fa0f89080443a970ae032cea4c204699d113", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/openai/universe/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:05Z", "role": "primary", "sha256": "ecd4a4aacc4512a31bbdc3252ffc014d37ac7ef56ed34bc2a5d0f6337df6daa5", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/openai/universe/HEAD/LICENSE"}], "id": "github:openai/universe", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-17T11:20:17Z", "licenses": {"code": {"basis": "Official Universe client-source LICENSE; no license or availability claim is transferred to external remote applications or container contents.", "declared": "MIT", "retrieved_at": "2026-10-17T11:23:05Z", "url": "https://raw.githubusercontent.com/openai/universe/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/openai/universe"}], "name": "Universe legacy GUI interaction client", "notes": "GitHub marks this repository archived, and the README declares the library deprecated. It contains the client side of Universe; the README says remote-environment source repositories would be released separately. Original MiniWoB task-server or container availability was not established: alpha.openai.com no longer resolved in this audit and the official Quay tag metadata endpoint returned 401. This entry records the released client infrastructure, not a currently runnable original MiniWoB benchmark. Legacy host requirements are documented; no container or code was executed.", "published_at": null, "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "An archived Gym-compatible client that exposes screen observations and keyboard/mouse actions over VNC for remote application and browser environments, including the historical World of Bits integration.", "topics": ["automation", "reinforcement-learning", "history"], "updated_at": null, "url": "https://github.com/openai/universe"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2cf79e21b5438bf6a2e6bd450ba8c4d86404560dfb61836f7e300112ce74a59d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/LICENSE"}], "id": "github:openbmb/agentcpm-gui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenBMB/AgentCPM-GUI"}], "name": "AgentCPM-GUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/README.md", "target_id": "arxiv:2506.01391", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "8B on-device smartphone agent built on MiniCPM-V that operates Chinese and English apps from screenshots, trained with reinforcement fine-tuning and a compact JSON action space.", "topics": ["grounding", "mobile-navigation"], "updated_at": null, "url": "https://github.com/OpenBMB/AgentCPM-GUI"}
{"aliases": [], "authors": [], "details": {"availability": "Public Android executor, agents, prompts and setup documentation.", "license_review": {"code": {"evidence_urls": ["https://api.github.com/repos/OpenDFM/MobA/git/trees/main?recursive=1"], "reason": "No repository source-license declaration established from the current public tree."}}, "repository_revision": "05cd237d80fcd23ad0dd6096cdb144117ecb3471"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "49c096e5ee6a87205497d693b5248817ea98cad0f7e5254793660007ef4292f7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/OpenDFM/MobA/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "0ecd1833a1b6840441769e1cc26a55268f2de7beb61bc3e783a851dd7cff45f1", "supports": ["identity"], "url": "https://api.github.com/repos/OpenDFM/MobA"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:04Z", "role": "primary", "sha256": "302f324fce2884c5948cf62c75f2f839cd40c301faff6ccf4b6eb5dd29f953a3", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/OpenDFM/MobA/git/trees/main?recursive=1"}], "id": "github:opendfm/moba", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T12:30:53Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenDFM/MobA"}], "name": "MobA", "notes": "The repository releases a configurable Python executor and Android ADB integration; its setup documents a Windows host, but the controlled interface is Android. The source is marked under continuing development and requires external model credentials. The inspected repository has no license file; the separately hosted MobBench data declaration is not transferred to this code.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenDFM/MobA/HEAD/README.md", "target_id": "arxiv:2410.13757", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An Android agent that combines global task planning and memory with a local action executor, using reflection to revise mobile interaction steps.", "topics": ["mobile-navigation", "memory", "planning", "self-reflection"], "updated_at": null, "url": "https://github.com/OpenDFM/MobA"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md", "population": "recommended hflqf88888/GUIOdyssey release", "qualifier": "exact", "unit": "episodes", "value": 8334}], "evidence_urls": ["https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md", "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md", "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/introduction.md"], "formats": ["JSON annotations", "PNG screenshots"], "modalities": ["screenshots", "instructions", "actions", "bounding boxes", "step context"], "notes": "The repository gives conflicting episode totals (8,334 and 8,834). This profile uses 8,334 from the specifically recommended current dataset card. No independent archive recount was performed. Coordinates are normalized to 0–1000.", "splits": ["random", "task", "device", "app"], "supervision": ["cross-app trajectories", "action grounding"]}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "4585b765d7c8b3e9bb6187d36d576edde454d4c0158a6aa8b3fdbedf34c89a15", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:39:52Z", "role": "primary", "sha256": "fd627ee0a6ee0596e0ee8147996b9a7a3436f754fc95ed63cf5f408d778b2831", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:39:52Z", "role": "primary", "sha256": "78c19ee712487af60df1913f11572958564e59a1a2cc663b1b30d971855d7a28", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/introduction.md"}], "id": "github:opengvlab/gui-odyssey", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenGVLab/GUI-Odyssey"}], "name": "GUI-Odyssey", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md", "target_id": "arxiv:2406.08451", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Human-demonstrated Android episodes for cross-app navigation, spanning six task types, 212 apps and about 1,400 app combinations, for training and evaluating agents.", "topics": ["cross-app", "data-generation"], "updated_at": null, "url": "https://github.com/OpenGVLab/GUI-Odyssey"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "0538701c860afedd5d63c3839039699d8e5a14977d5c339b187ed13ad666c8fa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/LICENSE"}], "id": "github:opengvlab/scalecua", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenGVLab/ScaleCUA"}], "name": "ScaleCUA", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/README.md", "target_id": "arxiv:2509.15221", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Computer-use model family trained with cross-platform interaction data.", "topics": ["data-generation", "planning"], "updated_at": null, "url": "https://github.com/OpenGVLab/ScaleCUA"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "033d2de815e11c208862c5c5601b78d58668b79cadd37a750cad9edd4dacabed", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenGVLab/ZeroGUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenGVLab/ZeroGUI/HEAD/LICENSE"}], "id": "github:opengvlab/zerogui", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/OpenGVLab/ZeroGUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenGVLab/ZeroGUI"}], "name": "ZeroGUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenGVLab/ZeroGUI/HEAD/README.md", "target_id": "arxiv:2505.23762", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Online RL framework for GUI agents in which VLMs generate the training tasks and judge success from trajectory screenshots, with OSWorld and AndroidLab checkpoints trained from UI-TARS-7B-DPO.", "topics": ["reinforcement-learning", "data-generation"], "updated_at": null, "url": "https://github.com/OpenGVLab/ZeroGUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "8f6c33487eca77769798677d32bce15335d0439f129c9a0413e12e0d9e70ab7a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenInterpreter/open-interpreter/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "d17f227e4df5da1600391338865ce0f3055211760a36688f816941d58232d8dc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenInterpreter/open-interpreter/HEAD/LICENSE"}], "id": "github:openinterpreter/open-interpreter", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/OpenInterpreter/open-interpreter/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenInterpreter/open-interpreter"}], "name": "Open Interpreter", "notes": "The repository has evolved substantially; computer-use functionality is documented in its Computer Use section. Do not infer parity with historical releases.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Assistant framework combining code execution and computer interaction.", "topics": ["tool-integration"], "updated_at": null, "url": "https://github.com/OpenInterpreter/open-interpreter"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["GUI gestures", "foreground-app tool calls in hybrid mode"], "counts": [{"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/bench_env/benchmark/official.json", "population": "official.json", "qualifier": "exact", "unit": "task definitions", "value": 215}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md", "population": "complete simulator installation", "qualifier": "exact", "unit": "simulated clients", "value": 63}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md", "population": "new clients in this release", "qualifier": "exact", "unit": "simulated clients", "value": 35}], "evaluation": "Programmatic checkers inspect changes in application state; both modes share task definitions and initial states.", "evidence_urls": ["https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md", "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/bench_env/benchmark/official.json"], "focus": ["task completion", "cross-application workflows", "GUI and app-native tool comparison"], "notes": "The task count was checked against unique task_id values in the published manifest. The 35 new clients are a subset of 63 total clients. This simulator runs in a browser; Android-emulator apps are a separate part of OpenMobile-2.", "observations": ["screenshots"], "setting": "interactive"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "d9f850ec196e715650e679cb81515adaa64ddc5a60a882bfe97b0f6a098ddc83", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:06Z", "role": "primary", "sha256": "9b385e9f6aa2f7169f616a76887df3b76df605cb2bea071d99ee7ca3f9b659cc", "supports": ["relations"], "url": "https://os-copilot.github.io/OpenMobile2-Home/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:50:11Z", "role": "primary", "sha256": "35c1641b349201c68917fbe4c4036814487678e9588d9b1c903af26917800316", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/bench_env/benchmark/official.json"}, {"method": "license_header", "retrieved_at": "2026-10-17T01:52:51Z", "role": "primary", "sha256": "7618cd7de039c264759ca8ca0b9130c6de63a4536dcf436baedd4cfbf639073e", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/LICENSE"}], "id": "github:openmobile2/mobilegym-plusplus", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T01:52:51Z", "url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenMobile2/MobileGym-plusplus"}], "name": "MobileGym++ / MobileGym++ Bench", "notes": "These clients simulate named commercial applications and are not those commercial products. The code and simulator data have distinct license files. No benchmark run was performed. The README describes a phone that runs in the browser, and the Android-emulator playground is a separate OpenMobile2 release, so the platform facet is web rather than a mobile operating system.", "published_at": null, "relations": [{"evidence_url": "https://os-copilot.github.io/OpenMobile2-Home/", "target_id": "arxiv:2610.14157", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Browser-rendered mobile simulation with resettable app state and a shared task suite for GUI-only and GUI-plus-tool agents.", "topics": ["mobile-navigation", "evaluation", "cross-app", "tool-integration"], "updated_at": null, "url": "https://github.com/OpenMobile2/MobileGym-plusplus"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "485279038266d9454df75e27a80a0a0d71c24e90700126d2dcfa0da921b1273d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T01:52:51Z", "role": "primary", "sha256": "b09fcaf4fbc265223c2479c410beef2b98c5e687c78ecedf3452aa12ffc465f1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/LICENSE"}], "id": "github:openmobile2/openmobile2-code", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T01:52:51Z", "url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenMobile2/OpenMobile2-Code"}], "name": "OpenMobile-2 code", "notes": "The README marks evaluation code, checkpoints, environments, and data as released; data-construction scripts and other resources remain pending. Each benchmark runtime is started separately. Code terms do not establish dataset permissions.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/README.md", "target_id": "arxiv:2610.14157", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Evaluation adapters and training guidance for OpenMobile-2 across Android emulators and browser-rendered mobile simulation.", "topics": ["mobile-navigation", "tool-integration", "evaluation"], "updated_at": null, "url": "https://github.com/OpenMobile2/OpenMobile2-Code"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "021cf14ec9482d1dab04dfeaadc3e39cdc405a284991cd540f5eeae7dea47a12", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/LICENSE"}], "id": "github:os-copilot/os-atlas", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Atlas"}], "name": "OS-Atlas", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/README.md", "target_id": "arxiv:2410.23218", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "OS-Atlas GUI action models, including grounding bases fine-tuned from InternVL2-4B and Qwen2-VL-7B-Instruct that accept any image size and output 0-1000 relative coordinates.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/OS-Copilot/OS-Atlas"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "b840368441695a54137216c2ba6bc99a8d35df75ad4196a4f0f0a1f5b5a70a9d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Copilot/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "3704700ddeaef20ec20cad39ed105ea518b316bc80c04dbcf2e69c3dc768a34f", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Copilot/HEAD/LICENSE"}], "id": "github:os-copilot/os-copilot", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Copilot/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Copilot"}], "name": "OS-Copilot", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Copilot/HEAD/README.md", "target_id": "arxiv:2402.07456", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Open-source library for generalist agents acting on an operating system's web, terminals, files and third-party apps, with the FRIDAY assistant, which takes custom tools and self-learns Excel tasks.", "topics": ["planning", "tool-integration"], "updated_at": null, "url": "https://github.com/OS-Copilot/OS-Copilot"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"], "formats": ["JSONL trajectories", "JSON transition triples"], "modalities": ["screenshots", "text", "actions"], "notes": "The repository links distinct AndroidControl-oriented, AndroidWorld-oriented, and web training files plus raw pre-state/action/post-state triples. These are training artifacts, not a documented train/test split, and related representations are not disjoint corpus totals. No numerical inventory is asserted from release names or linked archives.", "splits": [], "supervision": ["reverse task synthesis", "GUI action trajectories"]}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "430e495ab250395d666854bd18b0fc29b819cd784cd7730215ef65e461834d0d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "430e495ab250395d666854bd18b0fc29b819cd784cd7730215ef65e461834d0d", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["details.dataset_profile"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "github:os-copilot/os-genesis", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Genesis"}], "name": "OS-Genesis", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md", "target_id": "arxiv:2412.19723", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Interaction-driven pipeline that synthesizes mobile and web GUI agent trajectories without human supervision or predefined tasks, via reverse task synthesis and a trajectory reward model.", "topics": ["data-generation"], "updated_at": null, "url": "https://github.com/OS-Copilot/OS-Genesis"}
{"aliases": [], "authors": [], "details": {"availability": "Public task definitions, execution harness and an ungated VM-file manifest.", "benchmark_protocol": {"actions": ["actions executed by the OSWorld-derived DesktopEnv"], "counts": [{"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/all.json", "population": "evaluation_examples/all.json", "qualifier": "exact", "unit": "category/task references", "value": 416}], "evaluation": "Task configurations provide evaluators for application/system state, using an OSWorld-derived runtime.", "evidence_urls": ["https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/README.md", "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/all.json", "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/README.md"], "focus": ["breadth of desktop task coverage", "depth of desktop task execution"], "notes": "416 category/task pairs are unique, but only 412 bare ID strings are unique across categories. This is a manifest-reference count, not a claim of 416 semantically distinct tasks. Main runtime documentation specifies Ubuntu; the subdirectory README retains older Windows-only wording.", "observations": ["desktop observations from an Ubuntu virtual machine"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/LICENSE"], "reason": "Repository code LICENSE; no license is inferred for the VM image or bundled third-party applications."}}, "repository_revision": "191efd0e6766acab444a31685b437565e1a94ac2"}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:14Z", "role": "primary", "sha256": "fc9f3c8d9fffdff8b51e312b013be0d9fdfc10e57fdc2ec4a8a8d012c4b42475", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:14Z", "role": "primary", "sha256": "d6c91b4910e92f40cc694b4cb6467c1a20f42c0c14047b95a1e777aca0892677", "supports": ["identity"], "url": "https://api.github.com/repos/OS-Copilot/OS-Map"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:07Z", "role": "primary", "sha256": "41342157c53f28ad2729137491930225ac11da922e7b16762e84ac958a2e9d28", "supports": ["release_status"], "url": "https://api.github.com/repos/OS-Copilot/OS-Map/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:07Z", "role": "primary", "sha256": "1dd924135cbf0a85ce72d75025259adb7b849970d400af4ed854327a34e7ee41", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:17:22Z", "role": "primary", "sha256": "4199af4c8bc60451079d8a2e7598ab7386bd82b0d5c9ad7e6927d6cb8e03ad6f", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/all.json"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:17:22Z", "role": "primary", "sha256": "a6b60d5721cfdd26416e7d0ae3192360896427a97ccb3793812cc150e9cd7126", "supports": ["details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:16Z", "role": "primary", "sha256": "d2da837872b123f2927a069a020aeb14fb4b22148599c9f2e8b6689ee444ec88", "supports": ["release_status", "details"], "url": "https://huggingface.co/api/datasets/os-map/OSMAP"}], "id": "github:os-copilot/os-map", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T09:12:14Z", "licenses": {"code": {"basis": "Repository code LICENSE; no license is inferred for the VM image or bundled third-party applications.", "declared": "MIT", "retrieved_at": "2026-10-17T09:14:07Z", "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Map"}], "name": "OS-Map", "notes": "Public task manifests, agent runner and evaluator code are present, and the linked ungated Hub manifest lists Ubuntu0.zip and task files. The main setup supports x86 VMware Workstation and says ARM/Docker alternatives are still in development. An inherited evaluation-examples README still mentions Windows 10 and under-construction status; those statements are retained as a documentation inconsistency, not interpreted as verified Windows support. The manifest has 416 category/task references but 412 globally distinct ID strings.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/README.md", "target_id": "github:xlang-ai/osworld", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/README.md", "target_id": "arxiv:2507.19132", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An OSWorld-derived desktop benchmark with an Ubuntu virtual-machine environment and expanded task configurations for measuring breadth and depth of computer interaction.", "topics": ["evaluation", "long-horizon", "cross-app"], "updated_at": null, "url": "https://github.com/OS-Copilot/OS-Map"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "42bf4aea8369ac2909bf27bcb1078756c2c26567e2b46284c8f3a9e4c4f636c1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Sentinel/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Sentinel/HEAD/LICENSE"}], "id": "github:os-copilot/os-sentinel", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Sentinel/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Sentinel"}], "name": "OS-Sentinel", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Sentinel/HEAD/README.md", "target_id": "arxiv:2510.24411", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Safety-checking environment and benchmark for mobile agents on AndroidWorld that judges proposed actions with rule-based and VLM-based checks, evaluated on the MobileRisk trajectory dataset.", "topics": ["safety-security"], "updated_at": null, "url": "https://github.com/OS-Copilot/OS-Sentinel"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["PyAutoGUI desktop actions", "code/tool commands and answer/stop actions according to configured agent mode"], "counts": [{"basis": "computed", "evidence_url": "https://api.github.com/repos/OS-Copilot/ScienceBoard/git/trees/main?recursive=1", "population": "tasks/VM/**/*.json at tree 21b5eb0f; excludes tasks/Raw and tasks/Dual", "qualifier": "exact", "unit": "VM task definitions", "value": 169}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/verifier-audit.md", "population": "VM suite: Celestia, ChimeraX, GrassGIS, KAlgebra, Lean, TeXstudio", "qualifier": "exact", "unit": "application families", "value": 6}], "evaluation": "Application-specific verifiers inspect state, files, database values, equations, compilation or answer/stop conditions. The official audit records 242 verifier entries across its VM task suite. This profile covers the released VM task definitions; Docker smoke results and Raw/Dual alternatives are separate populations.", "evidence_urls": ["https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md", "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/verifier-audit.md", "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/known-issues.md", "https://api.github.com/repos/OS-Copilot/ScienceBoard/git/trees/main?recursive=1", "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/sci/base/task.py"], "focus": ["scientific-software desktop workflows", "GUI and code-assisted scientific tasks"], "notes": "Manifest count 169 matches the official verifier audit. Other Raw/Dual definitions can represent alternative execution modes of the same problems and are not added. Official Docker migration documents a 20-task acceptance subset, not a full-suite performance result; legacy VMware remains. External service sign-in/rate limits, amd64 binaries and ChimeraX redistribution terms limit portability. Only code MIT is established; VM, scientific data and bundled apps need their own terms.", "observations": ["desktop screenshots", "accessibility-tree observations", "set-of-marks with accessibility tree"], "setting": "interactive"}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "4c3fe93aa740dbf4a279a246ddd836f2590a2a36f39b820abe166e38cb484577", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "d276e11127abc5826bd1b2aa2caf543fcd45a8a60477774e8c8b1054067997ec", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/verifier-audit.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "4f90270617c0c1d58f523c0a7f2e4603fa67a49560d68e2f88cfa080d764313b", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/known-issues.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:07Z", "role": "primary", "sha256": "3398670a517a5d1d71f420dd866493d3f8a77853e3933fd4928103f9a8be1e4f", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/OS-Copilot/ScienceBoard/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:35:24Z", "role": "primary", "sha256": "d6a1481b453c074e34739fd29def56efa35a1dc10e856418f557f067e9ff08c5", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/sci/base/task.py"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "cbb1d28b515a39c294963d4bfc24077e669a483fc5ebba9b84a727afd4187426", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/LICENSE"}], "id": "github:os-copilot/scienceboard", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/ScienceBoard"}], "name": "ScienceBoard", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md", "target_id": "arxiv:2505.19897", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Scientific-workflow tasks for multimodal agents in desktop software such as ChimeraX, KAlgebra, Celestia, Grass GIS and Lean 4, with apps modified to expose state for evaluation.", "topics": ["evaluation", "science"], "updated_at": null, "url": "https://github.com/OS-Copilot/ScienceBoard"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:54Z", "role": "primary", "sha256": "61eb6109404f266fd22c6fa8ca58ef773a9d502567403f42edc096958ab2e430", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/EIA_against_webagent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:46Z", "role": "primary", "sha256": "953ff47dbbc0c0563d1ec2d245b06d1206f2f38c4ed2f0da728e027ae2e5a973", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/EIA_against_webagent/HEAD/LICENSE"}], "id": "github:osu-nlp-group/eia_against_webagent", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T10:59:54Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-18T11:03:46Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/EIA_against_webagent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/EIA_against_webagent"}], "name": "EIA", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/EIA_against_webagent/HEAD/README.md", "target_id": "arxiv:2409.11295", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Evaluation suite for Environmental Injection Attacks that embed invisible malicious HTML elements into benign websites to test privacy leakage in generalist web agents.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/EIA_against_webagent"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["CLICK", "TYPE", "SELECT"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md", "population": "train split", "qualifier": "exact", "unit": "task instances", "value": 1009}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md", "population": "cross-task test split", "qualifier": "exact", "unit": "task instances", "value": 252}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md", "population": "cross-website test split", "qualifier": "exact", "unit": "task instances", "value": 177}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md", "population": "cross-domain test split", "qualifier": "exact", "unit": "task instances", "value": 912}], "evaluation": "Offline candidate and action evaluation against recorded demonstrations; the repository distinguishes macro from micro averaging.", "evidence_urls": ["https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md"], "focus": ["web action prediction", "generalization to unseen sites and domains"], "notes": "These are reported split inventories, not online task-success measurements. Original Mind2Web, Multimodal-Mind2Web, and Online-Mind2Web have different inputs/protocols. The authors require macro-average metrics for comparison with the paper; test archives remain upstream.", "observations": ["task instructions", "HTML", "recorded action history"], "setting": "offline"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "157dbd840264ea29f958c0b7383d63783845c573b3e19a1f39b773be22e0b32d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "157dbd840264ea29f958c0b7383d63783845c573b3e19a1f39b773be22e0b32d", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "953ff47dbbc0c0563d1ec2d245b06d1206f2f38c4ed2f0da728e027ae2e5a973", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/LICENSE"}], "id": "github:osu-nlp-group/mind2web", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/Mind2Web"}], "name": "Mind2Web", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md", "target_id": "arxiv:2306.06070", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Human demonstrations and tasks for generalist web navigation research.", "topics": ["evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/Mind2Web"}
{"aliases": [], "authors": [], "details": {"leaderboard_url": "https://osu-nlp-group.github.io/Mind2Web-2/#leaderboard"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:18:21Z", "role": "primary", "sha256": "5cc72b992cfc11094dcf341c284d5baadc6c66a302fdfd474b27913fe1c91f6f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web-2/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T17:18:56Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web-2/HEAD/LICENSE"}], "id": "github:osu-nlp-group/mind2web-2", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T17:18:21Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T17:18:56Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web-2/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/Mind2Web-2"}], "name": "Mind2Web 2", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Long-horizon agentic web search benchmark whose Agent-as-a-Judge framework scores time-varying, citation-backed answers, with evaluation scripts for its dev and test sets.", "topics": ["evaluation", "web-navigation", "long-horizon"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/Mind2Web-2"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["web interactions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md", "population": "documented Online-Mind2Web suite", "qualifier": "exact", "unit": "tasks", "value": 300}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md", "population": "documented site coverage", "qualifier": "exact", "unit": "websites", "value": 136}], "evaluation": "Human evaluation and WebJudge-based evaluation are documented as distinct assessment paths.", "evidence_urls": ["https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md"], "focus": ["live-web task completion"], "notes": "Tasks are actively replaced when websites or requirements become invalid. Record the task revision and submission schema; a constant headline task count does not imply an unchanged test set.", "observations": [], "setting": "interactive"}, "leaderboard_url": "https://huggingface.co/spaces/osunlp/Online_Mind2Web_Leaderboard"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "49d7cc317fd2ba0df65abf65a0357d1fa2384e59019a96d4e24b784888599eaa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "93675f9d13697ae98889d703b8fae86a7200623e7ae8d844c9ee2200ad480583", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/LICENSE"}], "id": "github:osu-nlp-group/online-mind2web", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/Online-Mind2Web"}], "name": "Online Mind2Web", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "300 tasks on 136 live websites for evaluating web agents online, with outdated tasks replaced over time, human-reviewed submissions and the WebJudge automatic evaluator.", "topics": ["evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/Online-Mind2Web"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "782dea8635b417bc89d21489cacea215ee6878ba5a1d4b9e18ca7b2913545ab6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/RedTeamCUA/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "10a9bb268e611da40ae6edba6777917abb6e97d43dcb480e772ae8169b687e10", "supports": ["identity"], "url": "https://api.github.com/repos/OSU-NLP-Group/RedTeamCUA"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "712f1b7bc13b593cf23ecfab72cec33d1c256afe23a383e49d7d5b2996a0f55c", "supports": ["release_status"], "url": "https://api.github.com/repos/OSU-NLP-Group/RedTeamCUA/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/RedTeamCUA/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/RedTeamCUA/HEAD/LICENSE"}], "id": "github:osu-nlp-group/redteamcua", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:44:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T02:45:20Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/RedTeamCUA/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/RedTeamCUA"}], "name": "RedTeamCUA", "notes": "The release includes environment code, adversarial scenario configuration and RTC-Bench task-generation assets. The project declares Apache-2.0, while incorporated environments retain their licenses. It distinguishes full task runs from tests initialized at injection locations; those settings must not be treated as interchangeable results.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/RedTeamCUA/HEAD/README.md", "target_id": "arxiv:2505.21936", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A controlled testing framework that combines desktop virtual machines and replicated websites to evaluate computer-use agents against indirect prompt injection.", "topics": ["safety-security", "evaluation", "sandbox"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/RedTeamCUA"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "673b0e309540e57ec3b977689b33f89ef80e45ba1c0ffb05e7412e4862f6b99e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SeeAct/HEAD/README.md"}], "id": "github:osu-nlp-group/seeact", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/SeeAct"}], "name": "SeeAct", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/SeeAct/HEAD/README.md", "target_id": "arxiv:2401.01614", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Generalist web agent framework for large multimodal models such as GPT-4V, with a Playwright tool that runs agents on live websites, text-choice and set-of-mark grounding, and a crawler mode.", "topics": ["web-navigation", "grounding"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/SeeAct"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "e06dbc9d58e713a0019d5040aa21220de85266e80aacc7f2d320a0518c781496", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SkillWeaver/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "4dfa648e22431d1a4903eb1e11f432fbe8a0f39d24e2e3324c300285ffe458fd", "supports": ["identity"], "url": "https://api.github.com/repos/OSU-NLP-Group/SkillWeaver"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "7a1c94c67131a97516d92632c00a2a20d09226fec57bb09e77ff6d84a4e6e79a", "supports": ["release_status"], "url": "https://api.github.com/repos/OSU-NLP-Group/SkillWeaver/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SkillWeaver/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:21Z", "role": "primary", "sha256": "822a60969166874dd8a26273b8d749ef8c69c46137a0915ed7178b3265bd037f", "supports": ["relations"], "url": "https://osu-nlp-group.github.io/SkillWeaver/"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SkillWeaver/HEAD/LICENSE"}], "id": "github:osu-nlp-group/skillweaver", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:44:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T02:45:20Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SkillWeaver/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/SkillWeaver"}], "name": "SkillWeaver", "notes": "Exploration, task execution, browser integration and evaluation code are published under MIT. Synthesized skill files are included, but no trained model weights are implied. The source depends on external web environments and model services; no runtime or browser tasks were executed.", "published_at": null, "relations": [{"evidence_url": "https://osu-nlp-group.github.io/SkillWeaver/", "target_id": "arxiv:2504.07079", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A web-agent framework that explores sites, synthesizes reusable interaction APIs and builds a skill library for later tasks.", "topics": ["web-navigation", "tool-integration", "data-generation"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/SkillWeaver"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "72ca3f23114c3a08fefb724573acf10c9f0a1b3d4a50d811c98356264ecf6ffc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/LICENSE.txt"}], "id": "github:osu-nlp-group/uground", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/LICENSE.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/UGround"}], "name": "UGround", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md", "target_id": "arxiv:2410.05243", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Code, training data and weights for universal visual grounding of GUI elements from screenshots, including Qwen2-VL-based UGround-V1 at 2B, 7B and 72B and offline and online evaluation suites.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/UGround"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "6602cb260b52ae5b2b93db231a51a38d0463c78d8e71321620e05b4fa31caacb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebDreamer/HEAD/README.md"}], "id": "github:osu-nlp-group/webdreamer", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/WebDreamer"}], "name": "WebDreamer", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebDreamer/HEAD/README.md", "target_id": "arxiv:2411.06559", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "World-model and planning components that simulate web-action outcomes, with linked training data and model checkpoints.", "topics": ["world-models", "planning"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/WebDreamer"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:53Z", "role": "primary", "sha256": "be53561dcf69cb2295a26eaebf965eead10b3504fca30f71f73dc5e5728fc5b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebGuard/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:46Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebGuard/HEAD/LICENSE"}], "id": "github:osu-nlp-group/webguard", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T10:59:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-18T11:03:46Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebGuard/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/WebGuard"}], "name": "WebGuard", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebGuard/HEAD/README.md", "target_id": "arxiv:2507.14293", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Codebase, annotation toolkit, and action-level risk dataset for training and evaluating guardrail models that flag unsafe browser-agent actions before execution.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/OSU-NLP-Group/WebGuard"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "8a7f7199f7cf52cb31940bcea8ae797d4ab6dbd2029c3fc9066a17dde3d19095", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OthersideAI/self-operating-computer/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "7104ee457c53f2e8e12fd536850e9c2e4894e5e23100f0aa74c8fa5924cec6b2", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OthersideAI/self-operating-computer/HEAD/LICENSE"}], "id": "github:othersideai/self-operating-computer", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/OthersideAI/self-operating-computer/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OthersideAI/self-operating-computer"}], "name": "Self Operating Computer", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Framework, released in November 2023, in which a multimodal model views the screen and picks mouse and keyboard actions toward a goal, with backends such as GPT-4o, Claude 3, Gemini and LLaVA.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/OthersideAI/self-operating-computer"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:14:29Z", "role": "primary", "sha256": "75f99d9388aee3e0b124ebdb5f33e08fbe8ec999ce1e9eba5cd71c21e60522e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/PathOnAI/LiteWebAgent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T13:29:25Z", "role": "primary", "sha256": "ec6ef142e3900e9f5555009b825387e41da67a77a24ee72ef6963c99b065e6eb", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/PathOnAIOrg/LiteWebAgent/HEAD/LICENSE"}], "id": "github:pathonaiorg/litewebagent", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T13:14:29Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T13:29:25Z", "url": "https://raw.githubusercontent.com/PathOnAIOrg/LiteWebAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/PathOnAIOrg/LiteWebAgent"}], "name": "LiteWebAgent", "notes": "The repository is maintained by the PathOnAI.org research community; its README explicitly disclaims affiliation with a for-profit company.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/PathOnAI/LiteWebAgent/HEAD/README.md", "target_id": "arxiv:2503.02950", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Modular web-agent framework with browser control, planning, workflow memory, and tree-search components, plus web and extension interfaces.", "topics": ["web-navigation", "planning", "memory", "tool-integration"], "updated_at": null, "url": "https://github.com/PathOnAIOrg/LiteWebAgent"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:38.123692Z", "role": "primary", "sha256": "3226df7608039522072e59d459094fc45be9c5a6be7ea0d7ad76ddf9723a3aea", "supports": ["name", "kind", "summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:04:53.470789+00:00", "role": "primary", "sha256": "3d76d934e98641746085772b98ee17059749f980a616435498a7ad483bd4cec1", "supports": ["id", "url"], "url": "https://api.github.com/search/repositories?q=topic%3Agui-grounding+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:51:05.389053Z", "role": "primary", "sha256": "4fe6f65a21586dde0bf8b64f8c49e5c583b320bf5e0c2b8253baf2cd3fcfc5d2", "supports": ["id", "url", "notes"], "url": "https://github.com/PatrickStarL/gui-grounding-eval"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:32.022035Z", "role": "primary", "sha256": "c3eb1708b7d57dcde7478ca2f5063241b1ab567b2ba2ba28b9956b043da2b6c2", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/ssp_eval/metrics.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:35.461491Z", "role": "primary", "sha256": "66fa5940f0c4f6265cb162cb0596402848dd2a49fd48b07ee950dc866b8730a5", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/ssp_eval/geometry.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:38.931130Z", "role": "primary", "sha256": "f83bc0632a616e7b01c8a46b0e435733f0cc47b4a19f0c709b221eaee8a0db5f", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/ssp_eval/runner.py"}], "id": "github:patrickstarl/gui-grounding-eval", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-20T11:51:38.123692Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/PatrickStarL/gui-grounding-eval"}], "name": "GUI Grounding Evaluation Harness", "notes": "This is an independent harness using ScreenSpot-Pro, not a new dataset or verified leaderboard entry. The read code maps points back to original-image coordinates and scores point-in-box hits; documentation limits comparisons by prompt, model, subset and provider behavior. No OS support, code license or external data/weight license was established in this bounded review. Model calls, tests and reported results were not executed or reproduced.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "ScreenSpot-Pro evaluation harness with explicit coordinate transforms, prompt and crop-based inference comparisons, and per-sample grounding accuracy and cost records.", "topics": ["grounding", "evaluation", "efficiency"], "updated_at": null, "url": "https://github.com/PatrickStarL/gui-grounding-eval"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:05Z", "role": "primary", "sha256": "8682f83d71dd4d44772db3e531f2a9fcb168729211e0cc43493020a0773b0aba", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/pixiebrix/agent-browser-shield/HEAD/README.md"}], "id": "github:pixiebrix/agent-browser-shield", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/pixiebrix/agent-browser-shield"}], "name": "Agent Browser Shield", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Browser extension for filtering potentially unsafe agent input.", "topics": ["safety-security"], "updated_at": null, "url": "https://github.com/pixiebrix/agent-browser-shield"}
{"aliases": [], "authors": [], "details": {"availability": "Public agent and browser-environment source; legacy API compatibility is unverified.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/LICENSE"], "reason": "Repository code license; not a license for the remote language-model service."}}, "repository_revision": "31bf737922fa0337d03367a844de5cf1aff4ac52"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:12Z", "role": "primary", "sha256": "6a101c867546365d21b48527781bd3af324075929067fb465c6105872070f692", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:11Z", "role": "primary", "sha256": "b826fc1747a1bbe38a8df8da567a68f79936509c5325ec76a683706118c526f5", "supports": ["identity"], "url": "https://api.github.com/repos/posgnu/rci-agent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:11Z", "role": "primary", "sha256": "3dcb8cc06679490315f0f3636bb2795e2f717b77d4156ad0bf2429cc36e3273c", "supports": ["release_status"], "url": "https://api.github.com/repos/posgnu/rci-agent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:12Z", "role": "primary", "sha256": "ffa8b2a603abd9ede8eb686e87c4a433110f1be47d7cc98df4977b7fb2ee72fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/LICENSE"}], "id": "github:posgnu/rci-agent", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T09:14:12Z", "licenses": {"code": {"basis": "Repository code license; not a license for the remote language-model service.", "declared": "MIT", "retrieved_at": "2026-10-17T09:14:12Z", "url": "https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/posgnu/rci-agent"}], "name": "RCI Agent", "notes": "The public release includes the agent, prompting utilities and a bundled MiniWoB++ environment wrapper. Its README documents older hosted model names and an OpenAI-key setup; contemporary endpoint compatibility was not tested. The repository notes that its MiniWoB++ dependency is not officially supported on Windows. MIT applies to repository code; no new trained weights are claimed.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/README.md", "target_id": "arxiv:2303.17491", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A MiniWoB++ agent that uses explicit and implicit recursive criticism-and-improvement prompts to refine action plans and ground actions to browser state.", "topics": ["web-navigation", "planning", "self-reflection", "training-free"], "updated_at": null, "url": "https://github.com/posgnu/rci-agent"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:05Z", "role": "primary", "sha256": "a495bb4c86cf54cd5e2276433cfecb401fd133287406b1a7e04c3eaaac6305ac", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T01:52:51Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/LICENSE"}], "id": "github:prentisai/groundgui", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T01:46:05Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T01:52:51Z", "url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/PrentisAI/GroundGUI"}], "name": "GroundGUI training and evaluation code", "notes": "The repository documents separate training-stage environments. Code, model weights, and ScreenRef data have distinct license declarations; no training or benchmark execution was reproduced here.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Code for screenshot grounding through supervised training, teacher-guided distillation, reinforcement learning, and benchmark evaluation.", "topics": ["grounding", "reinforcement-learning", "data-generation"], "updated_at": null, "url": "https://github.com/PrentisAI/GroundGUI"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["search queries", "click actions", "product option selection"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md", "population": "crowd-sourced instruction collection", "qualifier": "exact", "unit": "instructions", "value": 12087}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md", "population": "full product inventory", "qualifier": "approximately", "unit": "products", "value": 1180000}], "evaluation": "", "evidence_urls": ["https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md"], "focus": ["instruction-conditioned product selection", "web navigation"], "notes": "This is a simulated shop. The default preview loads 1,000 products, while the full inventory is much larger; report the loaded subset and observation mode. Instructions and products are different populations. Detailed reward semantics are not normalized in this profile.", "observations": ["HTML pages", "simplified text views"], "setting": "interactive"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "79b7a90a1413a52deb53142b5bbd81ecf40f61743bd4601c2835e91375898ec0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "79b7a90a1413a52deb53142b5bbd81ecf40f61743bd4601c2835e91375898ec0", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:58Z", "role": "primary", "sha256": "8872dbf8660b00890b5e07ce1ea1f7a44fa7ae4e1857da56caba172b51dab3cb", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/LICENSE.md"}], "id": "github:princeton-nlp/webshop", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:58Z", "url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/LICENSE.md"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/princeton-nlp/WebShop"}], "name": "WebShop", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md", "target_id": "arxiv:2207.01206", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Simulated shopping environment for instruction-following web agents.", "topics": ["evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/princeton-nlp/WebShop"}
{"aliases": [], "authors": [], "details": {"availability": "Public simulator, task/judge and online-RL source, plus official release metadata for the companion data archive.", "benchmark_protocol": {"actions": ["tap", "type", "swipe", "back", "home", "drag", "complete"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md", "population": "publisher benchmark overview", "qualifier": "exact", "unit": "task templates", "value": 416}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md", "population": "publisher benchmark overview", "qualifier": "exact", "unit": "simulated apps", "value": 28}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/train.txt", "population": "bench_env/splits/train.txt", "qualifier": "exact", "unit": "task identifiers", "value": 160}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/test.txt", "population": "bench_env/splits/test.txt", "qualifier": "exact", "unit": "task identifiers", "value": 256}], "evaluation": "Deterministic task judges inspect state changes; reported metrics distinguish success, progress, false completion and unexpected side effects.", "evidence_urls": ["https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md", "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/train.txt", "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/test.txt"], "focus": ["mobile GUI task completion in simulation", "state verification and parallel training"], "notes": "The split-ID inventory is not a count of sampled trajectories, live Android apps or real user accounts. Extra runtime/data dependencies and third-party content terms remain distinct.", "observations": ["simulated mobile screenshots", "structured simulator state for judges"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE"], "reason": "Core source LICENSE; mobilegym-rl has a separate matching Apache-2.0 license and retains vendored notices."}, "data": {"evidence_urls": ["https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE-DATA"], "reason": "Limited to maintainers’ dataset compilation/curation and authored synthetic content; third-party user content remains subject to platform terms and its authors’ rights."}}, "release_tag": "v0.1.0", "repository_revision": "5fbd3f34be0e9780ae4b1e9d5089963ddbd05d38"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:19Z", "role": "primary", "sha256": "815294dd2aa4bf8264978c3bfedee85a3a2c835c511860c732314a4c282a974a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:18Z", "role": "primary", "sha256": "48df04a2fbf479f956d5c7f22d7006c22f591b8857cab8a1d64c196349045495", "supports": ["identity"], "url": "https://api.github.com/repos/Purewhiter/mobilegym"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:56Z", "role": "primary", "sha256": "68f1fa16c0999079bccc3aa1b6c3c22b72b570a5b971837c5026ea5408853375", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/Purewhiter/mobilegym/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:08Z", "role": "primary", "sha256": "cb1057b0a0a2ef7ab781414c168bb41f0f3da88afc042ce266d6063493561bde", "supports": ["release_status", "details.release_tag"], "url": "https://api.github.com/repos/Purewhiter/mobilegym/releases/tags/v0.1.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:09Z", "role": "primary", "sha256": "ad0c86a6cdde2caba8bd1f98b121803817d93718319a863b0b833da2760c23aa", "supports": ["release_status"], "url": "https://api.github.com/repos/Purewhiter/mobilegym/releases/tags/data-v0.1.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:59Z", "role": "primary", "sha256": "7618cd7de039c264759ca8ca0b9130c6de63a4536dcf436baedd4cfbf639073e", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:27:50Z", "role": "primary", "sha256": "58d1e17ffe5109a7ae296caafcadfdbe6a7d176f0bc4ab01e12a689b0499d8bd", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/mobilegym-rl/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:59Z", "role": "primary", "sha256": "e222fa518c5003d425ee9b371e987c95ba10e52b54fea078571254ae6a8eaf23", "supports": ["licenses.data", "notes"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE-DATA"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:27:49Z", "role": "primary", "sha256": "0ecef460fecfd8aa91dcbb2cd4cca17b033f620f85a50d7f0b598ed19a27c0eb", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/train.txt"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:27:48Z", "role": "primary", "sha256": "a3d485015d480f4f1ebbd12dbbf02c25c0aa667d1afcf5f6fa35f0263a82c567", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/test.txt"}], "id": "github:purewhiter/mobilegym", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T11:18:19Z", "licenses": {"code": {"basis": "Core source LICENSE; mobilegym-rl has a separate matching Apache-2.0 license and retains vendored notices.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T11:22:59Z", "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE"}, "data": {"basis": "Limited to maintainers’ dataset compilation/curation and authored synthetic content; third-party user content remains subject to platform terms and its authors’ rights.", "declared": "CC-BY-NC-4.0", "retrieved_at": "2026-10-17T11:22:59Z", "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE-DATA"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Purewhiter/mobilegym"}], "name": "MobileGym", "notes": "The inspected code release is v0.1.0, with a separately listed data-v0.1.0 archive. The interaction target is a browser-hosted Android-like simulation; real-device transfer results are separate publisher-reported experiments. Source split manifests contain 160 training IDs and 256 test IDs with no duplicates or overlap. Platform/training code declares Apache-2.0. LICENSE-DATA applies CC-BY-NC-4.0 only to maintainer-created compilation, curation and synthetic content, and explicitly leaves third-party UGC/API terms and author rights separate. No runtime, archive or model was executed/downloaded.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md", "target_id": "arxiv:2605.26114", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A browser-hosted mobile-app simulator with programmable state, parameterized tasks and deterministic state-checking judges for parallel GUI-agent evaluation and online reinforcement learning.", "topics": ["evaluation", "mobile-navigation", "reinforcement-learning", "sandbox"], "updated_at": null, "url": "https://github.com/Purewhiter/mobilegym"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/LICENSE"], "reason": "Complete root MIT license text inspected at the pinned commit. Applies to repository software; data, weights, external WebDriverAgent code and separately distributed products do not inherit this declaration."}}, "repository_revision": "738cecdd3eacb9df88d48efb9efa53507bb4aeb4"}, "environments": ["ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:57:18.994340+00:00", "role": "primary", "sha256": "dfffdc4b5850ed6bfe0bcfe705fa5918d61420f8c80df2e38f61b7d1fb3819ec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:06:00.276978+00:00", "role": "primary", "sha256": "01f97a863952373bd10a1f849d79706344be63f5774fe7e0cfe9ba774e41d213", "supports": ["identity"], "url": "https://api.github.com/search/repositories?q=topic%3Amobile-agent+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:54:17.260430+00:00", "role": "primary", "sha256": "fde31e2fdae142655f09fa14fb3114bf39a22cc1ff7994204a66622f5cbfe2c7", "supports": ["identity", "details.repository_revision"], "url": "https://api.github.com/repos/RadishSoftware/mobster/commits/main"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:56:37.457839+00:00", "role": "primary", "sha256": "8c2b9647b33a93b1a33e9256307e09b0f3e6aa62608019defa6dcdc5d7bff3dc", "supports": ["identity", "details.repository_revision", "notes"], "url": "https://api.github.com/repos/RadishSoftware/mobster/git/trees/81f2c1990c07a3aaf001a5c3eb4f37a457f0fb3d?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:57:25.539870+00:00", "role": "primary", "sha256": "f9ad743798f16491100ea692069e346f82bbdb5a5df91e4e01c1d6168f2513cf", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:05:06.711971+00:00", "role": "primary", "sha256": "4296a186cc2c00b9de93ababae490dc99cba596d858eff195e40283d58d73a0a", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/mobile_agent/drivers.py"}], "id": "github:radishsoftware/mobster", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-20T11:57:18.994340+00:00", "licenses": {"code": {"basis": "Complete root MIT license text inspected at the pinned commit. Applies to repository software; data, weights, external WebDriverAgent code and separately distributed products do not inherit this declaration.", "declared": "MIT", "retrieved_at": "2026-10-20T11:57:25.539870+00:00", "url": "https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/RadishSoftware/mobster"}], "name": "Mobster engine", "notes": "This entry covers the published engine, not the separate Mobster for Mac product. The documented reference example requires an already-serving WebDriverAgent runner for an iPhone and a model-service key; it performs a fixed read-only Settings task and declines approvals. The repository lists benchmark/evaluation tooling, but dated simulator subsets, real-device trials and earlier engine results are different evaluation populations. The README explicitly places several product modules outside this repository. Source inspection confirms WDA dispatch paths, not task success, security guarantees or reproducible performance; no phone, simulator, model or benchmark was run. MIT applies to the inspected repository software, not automatically to external WDA code, images or other data.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "iOS agent engine organized around an observe, decide, verify and act loop, using WebDriverAgent screen state and phone actions such as taps, typing, swipes and app switches.", "topics": ["mobile-navigation", "evaluation"], "updated_at": null, "url": "https://github.com/RadishSoftware/mobster"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "5fd1399c88bdf406e7689765ccf0599b3b7e0ddbed082fa92bffdd7b120bd41a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/LICENSE"}], "id": "github:reallm-labs/infigui-r1", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Reallm-Labs/InfiGUI-R1"}], "name": "InfiGUI-R1", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/README.md", "target_id": "arxiv:2504.14239", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "3B GUI agent trained in two stages: distilling spatial reasoning from teacher models, then reinforcement learning that rewards accurate sub-goals and trains error recovery.", "topics": ["reinforcement-learning", "planning"], "updated_at": null, "url": "https://github.com/Reallm-Labs/InfiGUI-R1"}
{"aliases": [], "authors": [], "details": {"availability": "Public browser-library source and npm package metadata.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/LICENSE"], "reason": "Publisher source license for WebMarker."}}, "package_version": "0.7.0", "repository_revision": "03cf632eef3f76bfbc4fd9c23b31baa0355e41c7"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:56Z", "role": "primary", "sha256": "b46edb85d41e033f84bdd9bb96122f801364448f28a36fa6deccc110a19b6ffc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:56Z", "role": "primary", "sha256": "a44253aa866acd38fc3e2275aae61ecd051f956b18cb7b29346c699cf6241d6b", "supports": ["identity"], "url": "https://api.github.com/repos/reidbarber/webmarker"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "fc8de18814136ca3684749363727f263908c23e6ae274a0ed589923774aa3a95", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/reidbarber/webmarker/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "1caae04ee16fcc72c48439677f5aba55128879cd99d6d55fd72caa85a7030e69", "supports": ["identity", "release_status", "details.package_version"], "url": "https://registry.npmjs.org/webmarker-js/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "62dcdb525872e3f49d90ff9b354b79341ac2574522bca7c7146f79040491a627", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/LICENSE"}], "id": "github:reidbarber/webmarker", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T12:30:56Z", "licenses": {"code": {"basis": "Publisher source license for WebMarker.", "declared": "MIT", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/reidbarber/webmarker"}], "name": "WebMarker", "notes": "Source and npm release metadata for webmarker-js 0.7.0 are public. The documented mark/unmark API supports screenshot prompts and Playwright interaction, but this utility is not itself a language model or autonomous agent. The WebVoyager/Set-of-Mark references are usage inspiration and are not asserted as papers introducing this library. No package was installed or executed.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A browser library that labels interactive elements and their boxes for visual prompting, preserving element mappings that automation code can use to execute selected actions.", "topics": ["grounding", "screen-understanding", "tool-integration"], "updated_at": null, "url": "https://github.com/reidbarber/webmarker"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "2a313a5a1c433c0609067c6d875d5ff69493e8792d2173e458eb5db8bf232977", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/reworkd/tarsier/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "f4ff2ba0f16e8abb7087071651c217ced6421e224313755c5d88433e324f19a1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/reworkd/tarsier/HEAD/LICENSE"}], "id": "github:reworkd/tarsier", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/reworkd/tarsier/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/reworkd/tarsier"}], "name": "Tarsier", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Perception utilities for web agents that tag interactable elements with typed IDs an LLM can act on, and use OCR to render page screenshots as whitespace-structured text for models without vision.", "topics": ["screen-understanding"], "updated_at": null, "url": "https://github.com/reworkd/tarsier"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "94c7c78bede11dde727940f56592a04594c1c59438812dff47540bf53c1bed7d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/LICENSE"}], "id": "github:ritzz-ai/gui-r1", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ritzz-ai/GUI-R1"}], "name": "GUI-R1", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md", "target_id": "arxiv:2504.10458", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI action-model training and evaluation code with verifiable rewards, plus linked reinforcement-learning data and checkpoints.", "topics": ["reinforcement-learning", "grounding"], "updated_at": null, "url": "https://github.com/ritzz-ai/GUI-R1"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android", "ios", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:55Z", "role": "primary", "sha256": "cb3d2329571a80b6edeab98c04f8b63c7f4e976929788d6a8283238b0b7fdadc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "notes"], "url": "https://raw.githubusercontent.com/runamu/monday/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:47Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/runamu/monday/HEAD/LICENSE"}], "id": "github:runamu/monday", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T10:59:55Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T11:03:47Z", "url": "https://raw.githubusercontent.com/runamu/monday/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/runamu/monday"}], "name": "MONDAY", "notes": "The retained README marks release of the full data-generation pipeline, training scripts, and model checkpoints as pending. Linked video content and datasets remain subject to their original licenses; the repository code license does not establish a blanket data or weights license.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/runamu/monday/HEAD/README.md", "target_id": "arxiv:2505.12632", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Cross-platform mobile GUI trajectory dataset derived from instructional videos, with released data-processing helpers and example evaluation code for Android and iOS data.", "topics": ["data-generation", "mobile-navigation", "video"], "updated_at": null, "url": "https://github.com/runamu/monday"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "5559ca58f8ee06d2812c6cbc8c3e49ef112f4ef41fdcf4e95568b268bb24bf14", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/SAAgent/MCPWorld/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["relations", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "5559ca58f8ee06d2812c6cbc8c3e49ef112f4ef41fdcf4e95568b268bb24bf14", "supports": ["notes"], "url": "https://raw.githubusercontent.com/SAAgent/MCPWorld/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "d7a04bf11ce24a7b79cd44b4a78dcc9971eaeb3b565821de437e0ecb7723e634", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/SAAgent/MCPWorld/HEAD/LICENSE"}], "id": "github:saagent/mcpworld", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/SAAgent/MCPWorld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/SAAgent/MCPWorld"}], "name": "MCPWorld", "notes": "The retained paper reports 201 tasks, whereas the retained repository README describes approximately 170. These are separate source descriptions; their release/version difference has not been established.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2506.07672", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Evaluation environment combining computer interaction with tool access.", "topics": ["evaluation", "tool-integration"], "updated_at": null, "url": "https://github.com/SAAgent/MCPWorld"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "24251854f9481af5c8db65750197604546f000b54868930a6de79adc365db854", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/SALT-NLP/PopupAttack/HEAD/README.md"}], "id": "github:salt-nlp/popupattack", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/SALT-NLP/PopupAttack"}], "name": "PopupAttack", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Resources for evaluating GUI agents under malicious popup instructions.", "topics": ["safety-security"], "updated_at": null, "url": "https://github.com/SALT-NLP/PopupAttack"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/README.md", "https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/LICENSE"], "reason": "The retained README states that the project is MIT licensed, while the retained LICENSE contains Creative Commons Attribution-NonCommercial-NoDerivatives 4.0. These declarations conflict, so a settled code license is not asserted."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:54Z", "role": "primary", "sha256": "a9d33536fae38485eb6aac9698ae22ed3533510d43f783d6c0e9f9cf95d9c58f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.code", "notes"], "url": "https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:46Z", "role": "primary", "sha256": "f7f28b8c7a1af76b9874ca6d040e8b1eb6768fd1043106c9d315090ea96754e7", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:46Z", "role": "primary", "sha256": "f7f28b8c7a1af76b9874ca6d040e8b1eb6768fd1043106c9d315090ea96754e7", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/LICENSE"}], "id": "github:scaleapi/browser-art", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T10:59:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/scaleapi/browser-art"}], "name": "BrowserART", "notes": "The retained README and LICENSE contain conflicting code-license declarations. The catalog leaves the code-license value unknown pending an upstream clarification; it does not infer data or model permissions.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/README.md", "target_id": "arxiv:2410.13886", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Red-teaming benchmark and test suite from Scale AI evaluating whether refusal-trained language models can be jailbroken when deployed as browser agents.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/scaleapi/browser-art"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:54Z", "role": "primary", "sha256": "e712feb15886d5a2aba1051d419dc69dc6e58c9a86a9d2d566090bbbf67492fc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/segev-shlomov/ST-WebAgentBench/HEAD/README.md"}], "id": "github:segev-shlomov/st-webagentbench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T10:59:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/segev-shlomov/ST-WebAgentBench"}], "name": "ST-WebAgentBench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/segev-shlomov/ST-WebAgentBench/HEAD/README.md", "target_id": "arxiv:2410.06703", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "BrowserGym-based benchmark for evaluating the safety and policy compliance of enterprise web agents across six trustworthiness risk dimensions.", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/segev-shlomov/ST-WebAgentBench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "ca312434db1db852f96fb49bbc0a61f6a93cffe6c0dd3aa2b7ac814681ccc19e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ServiceNow/AgentLab/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "b192c58991e8ff585cc574615d40e74185404d4b96c1109d423071ab1367344b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ServiceNow/AgentLab/HEAD/LICENSE"}], "id": "github:servicenow/agentlab", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/ServiceNow/AgentLab/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ServiceNow/AgentLab"}], "name": "AgentLab", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ServiceNow/AgentLab/HEAD/README.md", "target_id": "arxiv:2412.05467", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Experiment framework for developing and evaluating browser agents.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/ServiceNow/AgentLab"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["element-ID actions", "coordinate mouse and keyboard actions", "tab/navigation/scroll operations", "Python/Playwright action mode"], "counts": [], "evaluation": "Gymnasium reset/step interface delegates reward, completion, messages and task info to each adapter’s validate() method. There is no universal BrowserGym success metric or shared task total; retain benchmark/adapter version and action subset for comparisons.", "evidence_urls": ["https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/README.md", "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/browsergym/core/src/browsergym/core/env.py", "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/docs/src/core/action_space.rst", "https://api.github.com/repos/ServiceNow/BrowserGym/git/trees/main?recursive=1"], "focus": ["shared browser-agent evaluation interface", "benchmark adapters with separate task definitions"], "notes": "A framework with interactive browser adapters and a static WebLINX evaluation integration, so the profile setting is mixed. README-listed integrations are not added as disjoint tasks or datasets. Core requires Playwright/Chromium; individual benchmarks have their own setup, credentials, licenses and validation rules. Apache-2.0 covers repository code only. No browser tasks or third-party project code were executed.", "observations": ["browser screenshot", "DOM snapshot", "accessibility tree", "tab/URL state", "previous action and its error", "chat messages and task goal"], "setting": "mixed"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "7b685115ef822989fac9093518d16c46f0202736d005eed56e7f9fc36b43f3ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:05Z", "role": "primary", "sha256": "867cc6efdfead356839fef8a26b492502b8b5e3d165d1b3c17d24aaeec268d63", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/browsergym/core/src/browsergym/core/env.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:11Z", "role": "primary", "sha256": "c4bbe6b7d0cb36a5d5effab1aa5568301cead97fe4015273e1b82c667da3823a", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/docs/src/core/action_space.rst"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:08Z", "role": "primary", "sha256": "418efbe8c65ad7ebab61d3a066339876c820a9c957506eb456b8078ab352f6bf", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/ServiceNow/BrowserGym/git/trees/main?recursive=1"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "b192c58991e8ff585cc574615d40e74185404d4b96c1109d423071ab1367344b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/LICENSE"}], "id": "github:servicenow/browsergym", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ServiceNow/BrowserGym"}], "name": "BrowserGym", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/README.md", "target_id": "arxiv:2412.05467", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Unified interfaces for browser-agent tasks and benchmark environments.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/ServiceNow/BrowserGym"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["BrowserGym browser actions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "population": "WorkArena-L1 atomic tasks", "qualifier": "exact", "unit": "task definitions", "value": 33}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "population": "WorkArena-L1 documented instances", "qualifier": "exact", "unit": "task instances", "value": 19912}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "population": "WorkArena++ compositional tasks", "qualifier": "exact", "unit": "task definitions", "value": 682}], "evaluation": "Task-specific checks on ServiceNow workflows; levels and task configurations define different evaluations.", "evidence_urls": ["https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"], "focus": ["enterprise web workflows", "compositional tasks"], "notes": "Definitions and instantiated tasks are different units. ServiceNow instance access is gated. WorkArena-L1 and WorkArena++ should be reported separately.", "observations": ["multimodal browser observations"], "setting": "interactive"}, "leaderboard_url": "https://huggingface.co/spaces/ServiceNow/browsergym-leaderboard"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "834a48f0346036a0fc591a36f9eb6f513b44a2777a66afa0917f28fc63f29334", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "b192c58991e8ff585cc574615d40e74185404d4b96c1109d423071ab1367344b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/LICENSE"}], "id": "github:servicenow/workarena", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ServiceNow/WorkArena"}], "name": "WorkArena", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "target_id": "arxiv:2403.07718", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "target_id": "arxiv:2407.05291", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Browser-based enterprise workflow tasks built around ServiceNow.", "topics": ["evaluation", "workflow"], "updated_at": null, "url": "https://github.com/ServiceNow/WorkArena"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "9c43f97560c3781b28f97c894ff65bbea7121c3df8c183ddf15bb38842bd1e39", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/assistgui/HEAD/README.md"}], "id": "github:showlab/assistgui", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/assistgui"}], "name": "AssistGUI", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "LLM-based assistant for automating work in desktop productivity software by watching the screen and issuing mouse and keyboard commands; its repository says the code will be released later.", "topics": ["planning"], "updated_at": null, "url": "https://github.com/showlab/assistgui"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "7b53e58d3efb35d23e58bd75bfa6d36f1e57af9f63556c34935726cead47e322", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/computer_use_ootb/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "cebb89bae373fa4f9a210ed5b664fbbfcbdb7ac8b5d8c34bfa7e8905d1f63b4e", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/showlab/computer_use_ootb/HEAD/LICENSE"}], "id": "github:showlab/computer_use_ootb", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/showlab/computer_use_ootb/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/computer_use_ootb"}], "name": "Computer Use OOTB", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2411.10323", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Gradio desktop GUI agent for Windows and macOS that needs no Docker, using Claude 3.5 Computer Use over the API or ShowUI and UI-TARS run locally, with multi-display and phone remote control.", "topics": ["workflow"], "updated_at": null, "url": "https://github.com/showlab/computer_use_ootb"}
{"aliases": ["GUI-Narrator"], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50", "population": "Act2Cap benchmark described in the paper", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2406.13719", "unit": "video-captioning samples", "value": 4189}], "evidence_urls": ["https://github.com/showlab/GUI-Action-Narrator", "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"], "formats": [], "modalities": ["GUI video", "action captions"], "notes": "A captioning sample is not necessarily a unique video or complete task trajectory. Package layout and split sizes were not established by the inspected sources.", "splits": [], "supervision": ["GUI action captioning"]}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:52Z", "role": "primary", "sha256": "379aae1a8890a0c3d3b5999589b8338338cf1ea63883c2b77df678fd3bb367a2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://github.com/showlab/GUI-Action-Narrator"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["details.dataset_profile"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "id": "github:showlab/gui-action-narrator", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:40:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/GUI-Action-Narrator"}], "name": "GUI-Action Narrator", "notes": "Original repository URL redirects to showlab/GUI-Narrator.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Action narration resources for GUI interaction trajectories.", "topics": ["data-generation", "video"], "updated_at": null, "url": "https://github.com/showlab/GUI-Action-Narrator"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "730e51422c0ebe00cb2a1a8f4510d9914b1ea735fdb06445fde2b243da80bdd4", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "1b4977f976211c2644742e3a56cf75ecffb4c7c31d1bce456875123ba8bcd342", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/LICENSE"}], "id": "github:showlab/showui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/ShowUI"}], "name": "ShowUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/README.md", "target_id": "arxiv:2411.17465", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Lightweight GUI-agent vision-language-action model from Show Lab and Microsoft, released with UI-guided token selection, training code, the ShowUI-web and ShowUI-desktop datasets and vLLM inference.", "topics": ["grounding", "efficiency"], "updated_at": null, "url": "https://github.com/showlab/ShowUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "ad5553270b67c335df4d4e909f1549f43af8e74ad687f9cbcc812761affcad5c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/ShowUI-Aloha/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T11:15:55Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/showlab/ShowUI-Aloha/HEAD/LICENSE"}], "id": "github:showlab/showui-aloha", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T11:15:55Z", "url": "https://raw.githubusercontent.com/showlab/ShowUI-Aloha/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/ShowUI-Aloha"}], "name": "ShowUI-Aloha", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/ShowUI-Aloha/HEAD/README.md", "target_id": "arxiv:2601.07181", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Desktop agent that records demonstrations, derives semantic action traces, and uses them to plan and execute related workflows.", "topics": ["recording", "workflow", "automation"], "updated_at": null, "url": "https://github.com/showlab/ShowUI-Aloha"}
{"aliases": ["ShowUI-pi", "ShowUI Pi"], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "88df552ccbe80efd7020905d77d526f4f9efe2d09a11fed2714b083988f9dcef", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md"}], "id": "github:showlab/showui-pi", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/showui-pi"}], "name": "ShowUI-π", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md", "target_id": "arxiv:2512.24965", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Vision-language-action model and training code for GUI dragging, with links to checkpoints and continuous-action evaluation data.", "topics": ["grounding", "automation"], "updated_at": null, "url": "https://github.com/showlab/showui-pi"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "1a21631fb312cc7eac35547792611a84f13dcf756a3cf92209676ccdb92cbc6b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/videogui/HEAD/README.md"}], "id": "github:showlab/videogui", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/videogui"}], "name": "VideoGUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/videogui/HEAD/README.md", "target_id": "arxiv:2406.10227", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Benchmark for desktop GUI tasks derived from instructional videos.", "topics": ["evaluation", "video"], "updated_at": null, "url": "https://github.com/showlab/videogui"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "ea54ea82b86548ff38913ef328ec18152ace71a3fe9bb497bcabcaad24e6a678", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/WorldGUI/HEAD/README.md"}], "id": "github:showlab/worldgui", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/WorldGUI"}], "name": "WorldGUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/WorldGUI/HEAD/README.md", "target_id": "arxiv:2502.08047", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "611 desktop GUI tasks that agents must complete from varied starting points, shipped with instructional videos and project files, plus the self-reflective WorldGUI-Agent baseline.", "topics": ["evaluation", "robustness"], "updated_at": null, "url": "https://github.com/showlab/WorldGUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "abc848e7eb5181be22168dbc076446c5eb0f065fd57c7e10367284ca6e50bf5c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/shulin16/MMInA/HEAD/README.md"}], "id": "github:shulin16/mmina", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/shulin16/MMInA"}], "name": "MMInA", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/shulin16/MMInA/HEAD/README.md", "target_id": "arxiv:2404.09992", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Multimodal web tasks in a WebArena-based environment that chain from 1 to 10 hops across sites such as a shopping store and Wikipedia, testing long-chain reasoning.", "topics": ["evaluation", "long-horizon"], "updated_at": null, "url": "https://github.com/shulin16/MMInA"}
{"aliases": [], "authors": [], "details": {"availability": "Public legacy source and original documentation; maintainers explicitly state it is no longer supported.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/sikuli/sikuli/HEAD/copyright.txt"], "reason": "Original Sikuli copyright/license file for project source; bundled third-party libraries are not relicensed by this assertion."}}, "maintenance_status": "unsupported legacy research code", "repository_revision": "4adaab7880d2f3e14702ca7287ae9c9e4f4de9ab"}, "environments": ["desktop", "macos", "linux", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:12Z", "role": "primary", "sha256": "e752b2226af679a778e21506e660742376be3d7effe621f63c0b8816d3323729", "supports": ["name", "kind", "summary", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/index.rst"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:12Z", "role": "primary", "sha256": "90d840ab98d35b90c0f91f93dc2dc752eac631a4bdc299601bffe8856610dd4c", "supports": ["identity"], "url": "https://api.github.com/repos/sikuli/sikuli"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:49Z", "role": "primary", "sha256": "8ee5f1e75659bfbce0d9b22967ff2b51335be5f19a3dc9e645273cc51333153e", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/sikuli/sikuli/git/trees/develop?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:13Z", "role": "primary", "sha256": "ed7659fff11dac079b99ad3c264f7e1e0a97f313a75379067e1c7fccca23ef88", "supports": ["notes", "details.availability", "relations"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:13Z", "role": "primary", "sha256": "bab2f2dae13092e9c94f5d4b077bc96a6aa7da5257b54089f76ff5c0b0f61402", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/tutorials/helloworld/helloworld-mac.rst"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:13Z", "role": "primary", "sha256": "28354a06091997d233cc5124958d85b75a5d815677c58b4fbabd492a6b0038d3", "supports": ["environments"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/build-linux.txt"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:13Z", "role": "primary", "sha256": "b7c66064d08c6a768a4ed93fee0f86f837b41d51e60785d44a59d7a13bd2452a", "supports": ["environments"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/build-win32.txt"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:03Z", "role": "primary", "sha256": "717e5c54ef8201b2307c9828a6bacb7710385e5014adba93e5f1dc444339fdab", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/copyright.txt"}], "id": "github:sikuli/sikuli", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T11:28:12Z", "licenses": {"code": {"basis": "Original Sikuli copyright/license file for project source; bundled third-party libraries are not relicensed by this assertion.", "declared": "MIT", "retrieved_at": "2026-10-17T11:23:03Z", "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/copyright.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/sikuli/sikuli"}], "name": "Sikuli legacy research implementation", "notes": "This is the retained Sikuli X-1.0rc3 source line, explicitly marked unsupported by its maintainers. Its documented build targets include historical macOS, Linux and Windows setups; present-day build compatibility was not tested. The official documentation cites the 2009 search/automation and 2010 visual-testing papers. Follow-up SikuliX and current OculiX are distinguished from this legacy research implementation. MIT applies to project code; external libraries keep their own terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/index.rst", "target_id": "url:a4cdb2ee0d07d3980651", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/index.rst", "target_id": "url:6262c757080f8838e62c", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The original Sikuli source distribution combines screenshot targets with a Jython scripting library and IDE for visual GUI automation and testing.", "topics": ["automation", "grounding", "screen-understanding", "history"], "updated_at": null, "url": "https://github.com/sikuli/sikuli"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "9ea381fdb78eb326d988ea384b6f5074ef0ff4c0d5fab555a148eded385e88f1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "43070e2d4e532684de521b885f385d0841030efa2b1a20bafb76133a5e1379c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/LICENSE"}], "id": "github:simular-ai/agent-s", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/simular-ai/Agent-S"}], "name": "Agent S", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/README.md", "target_id": "arxiv:2410.08164", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/README.md", "target_id": "arxiv:2504.00906", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/README.md", "target_id": "organization:simular", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Simular's open-source framework for computer-use agents that click, type and scroll through desktop and web apps on macOS, Windows and Linux; its third generation, Agent S3, reports 72.6% on OSWorld.", "topics": ["planning", "memory"], "updated_at": null, "url": "https://github.com/simular-ai/Agent-S"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "68fda945255825907aeada13d19de19209708c14504ac975e63a235f0cd9343b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Skyvern-AI/skyvern/HEAD/README.md"}], "id": "github:skyvern-ai/skyvern", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Skyvern-AI/skyvern"}], "name": "Skyvern", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Skyvern-AI/skyvern/HEAD/README.md", "target_id": "organization:skyvern", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Browser workflow automation that relies on vision LLMs and a swarm of agents instead of fixed XPath selectors, offered as a Playwright-compatible SDK and a no-code workflow builder.", "topics": ["workflow"], "updated_at": null, "url": "https://github.com/Skyvern-AI/skyvern"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:56Z", "role": "primary", "sha256": "5653db27f736c0c77afb63f81ca051bf31f1e1f0579ac77cf00dca4ea751c159", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:47Z", "role": "primary", "sha256": "8486a10c4393cee1c25392769ddd3b2d6c242d6ec7928e1414efff7dfb2f07ef", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/LICENSE"}], "id": "github:skyworkai/agent-studio", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T10:59:56Z", "licenses": {"code": {"basis": "repository LICENSE header; GNU text does not say -only or -or-later; not a weights/data license", "declared": "AGPL-3.0", "retrieved_at": "2026-10-18T11:03:47Z", "url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/skyworkai/agent-studio"}], "name": "AgentStudio", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/README.md", "target_id": "arxiv:2403.17918", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Toolkit and Ubuntu-based environment for building and evaluating generalist computer-use agents through GUI and API interaction.", "topics": ["sandbox", "evaluation", "grounding", "recording"], "updated_at": null, "url": "https://github.com/skyworkai/agent-studio"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/LICENSE"], "reason": "Complete root MIT license text inspected at the pinned commit. Repository software only; data, weights and third-party assets retain separate terms."}}, "repository_revision": "2aa6cd83165bfc201b0128c250e80892895d63af"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:57:12.431753+00:00", "role": "primary", "sha256": "eea31b35aa0b55c97f8ccda3438382ad49d2453671b49bb817efb5bc3384e900", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:06:00.276978+00:00", "role": "primary", "sha256": "01f97a863952373bd10a1f849d79706344be63f5774fe7e0cfe9ba774e41d213", "supports": ["identity"], "url": "https://api.github.com/search/repositories?q=topic%3Amobile-agent+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:54:02.419278+00:00", "role": "primary", "sha256": "e2797712132e69cc50fa7ad2ae815a48339f77eebec1e85a5dc0cd75702574f1", "supports": ["identity", "details.repository_revision"], "url": "https://api.github.com/repos/solplum/yunkai/commits/main"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:56:30.226860+00:00", "role": "primary", "sha256": "08f48cb5c15b8d505e76a91d996c656d87765946cc7137bce088e73e6e64cf06", "supports": ["identity", "details.repository_revision", "notes"], "url": "https://api.github.com/repos/solplum/yunkai/git/trees/4ca94f2f7d8f13099ae2c351d9578ed5ef0cb509?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:57:22.238029+00:00", "role": "primary", "sha256": "c39441f92c303832d42f1f153c4442533ac925007e12d70f3199c568a0baac88", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:59:51.424713+00:00", "role": "primary", "sha256": "964aad706676fc7b74e532f8bc441fa766f3ce0ca3e4d0e115c7a90c70656883", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/yunkai-android/app/src/main/java/com/zhuolin/yunkai/service/screen/ScreenSenseService.kt"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:59:54.745065+00:00", "role": "primary", "sha256": "628cc3015999176052a09e3f3bbb8635d87163a9430d98cb6a6a7c09117dfd37", "supports": ["notes"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/yunkai-harmony/entry/src/main/ets/service/tools/M2Tools.ets"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:04:53.632582+00:00", "role": "primary", "sha256": "ba8c9b0f4b5cea5ac36a4b49f3cf27e77e5c6beb49022e43c667e359be2c6fca", "supports": ["environments", "notes"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/yunkai-harmony/entry/src/main/ets/service/screen/ScreenSenseExtension.ets"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:04:56.914360+00:00", "role": "primary", "sha256": "423a77e1fd57f3e105898f1adfc94c493de5c0de97fae4379f9cf540c89d4640", "supports": ["notes"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/yunkai-harmony/tests/fullflow/drive_harmony.sh"}], "id": "github:solplum/yunkai", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-20T11:57:12.431753+00:00", "licenses": {"code": {"basis": "Complete root MIT license text inspected at the pinned commit. Repository software only; data, weights and third-party assets retain separate terms.", "declared": "MIT", "retrieved_at": "2026-10-20T11:57:22.238029+00:00", "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/solplum/yunkai"}], "name": "Yunkai", "notes": "The Android AccessibilityService contains concrete screen-reading and action paths. The inspected HarmonyOS accessibility extension is a lifecycle/event scaffold with its screen-content probe disabled; HDC appears in an external UI-test driver, so Harmony GUI-control parity is not established by this review. Setup requires a configured OpenAI-compatible model endpoint and API key: local storage does not establish offline inference or a no-transmission guarantee. Documented approval and privacy mechanisms are not independently tested security guarantees. No app, device test or model inference was run.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Android phone assistant combining model-driven tool use with accessibility-tree reading, screenshots, and tap, swipe and text-entry primitives. Its documented interaction flow presents operation plans for user confirmation.", "topics": ["mobile-navigation", "screen-understanding", "human-in-the-loop"], "updated_at": null, "url": "https://github.com/solplum/yunkai"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "feee31aa4d0801cb364578099a13e4f9cdcaa4eb5fbf4312b3fc3f8edd479a14", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "eada6ee9aa180f55daca059c5e7a4922b320cccb0514864108eba9040b52415c", "supports": ["identity"], "url": "https://api.github.com/repos/SqueezeAILab/plan-and-act"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "10ff0a1c1abe25454541949fdaac1869d68b5d4426fbb5ef1d487657afcd26f5", "supports": ["release_status"], "url": "https://api.github.com/repos/SqueezeAILab/plan-and-act/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:15Z", "role": "primary", "sha256": "f86e78202a51ab013e1cc47ba37386134601317f5a0c4e4ed287981b52159122", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:02:15Z", "role": "primary", "sha256": "f86e78202a51ab013e1cc47ba37386134601317f5a0c4e4ed287981b52159122", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/LICENSE"}], "id": "github:squeezeailab/plan-and-act", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T02:02:15Z", "url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/SqueezeAILab/plan-and-act"}], "name": "Plan-and-Act", "notes": "Runnable entry points, framework modules and synthetic-data tooling are published. The repository code declares MIT; separate model and dataset cards have their own declarations. Its setup targets WebArena-Lite and requires that external environment.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/README.md", "target_id": "arxiv:2503.09572", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A web-agent framework that separates high-level planning from action execution, with replanning and synthetic planner-data generation components.", "topics": ["planning", "web-navigation", "data-generation"], "updated_at": null, "url": "https://github.com/SqueezeAILab/plan-and-act"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/sqzhang-lazy/D-PoT/HEAD/README.md"], "reason": "No project-wide code license is established by the inspected README/API/tree; third-party code presence does not establish a license for this package."}}, "repository_tree_sha": "fc764f19802f4f59afa635bfeec5a50e4c3d5465", "repository_url": "https://github.com/sqzhang-lazy/D-PoT"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "c9c21102db512270079666531dbab71c1ab2ec2073add16be56f8ab3c6fe8abf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.code"], "url": "https://raw.githubusercontent.com/sqzhang-lazy/D-PoT/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:02Z", "role": "primary", "sha256": "d62cfecbfccd120e53a7fc60e8d4b90b157e998cc5ddac467514c30c0cae1363", "supports": ["identity"], "url": "https://api.github.com/repos/sqzhang-lazy/D-PoT"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:24Z", "role": "primary", "sha256": "f75444a0252df5d6ec1ab70d805584cb490cdc38efa995aae55ffbde43ab5fb1", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/sqzhang-lazy/D-PoT/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "id": "github:sqzhang-lazy/d-pot", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:14Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/sqzhang-lazy/D-PoT"}], "name": "D-PoT evaluation code", "notes": "README describes AITW test sampling and preprocessing reused from Auto-GUI, followed by prediction/action-extraction/evaluation scripts. This is not claimed as a ready-to-run interactive phone controller. No project-level license declaration appears in the inspected README/API/non-truncated tree; code, dataset and model rights remain unestablished here.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50", "target_id": "arxiv:2410.00467", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Offline prompting and evaluation code for Dynamic Planning of Thoughts on Android-in-the-Wild, including action extraction and comparison with reference actions.", "topics": ["planning", "action-prediction", "evaluation"], "updated_at": null, "url": "https://github.com/sqzhang-lazy/D-PoT"}
{"aliases": [], "authors": [], "details": {"availability": "Public historical training and environment-wrapper source; legacy dependencies documented.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/LICENSE"], "reason": "Stanford WGE source license; not a blanket license for all third-party tasks, embeddings or demonstration resources."}}, "repository_revision": "bed526657a62a64a163f80deb4a8627cb63871a3"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:16Z", "role": "primary", "sha256": "50ee5234bf62320d096fbe4c11c7be21f50c9e09dae97754a16a5bbde2f1ab19", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:15Z", "role": "primary", "sha256": "66fcd6577d1867abf84e62fe9edf2e3f223292724cbe6b9d0ef02b582db5c8b2", "supports": ["identity"], "url": "https://api.github.com/repos/stanfordnlp/wge"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:50Z", "role": "primary", "sha256": "926aff11e14c529d05d9bc51fe12aadb007818119da55f99c2fa47d0387f466d", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/stanfordnlp/wge/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:17Z", "role": "primary", "sha256": "249b23ff62e21bd446b3e411b1d7ea7f173c7e19e42ed75e97800ed32201bb3a", "supports": ["details", "notes"], "url": "https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/third-party/miniwob-sandbox/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:05Z", "role": "primary", "sha256": "3df3dd1e3149cd6cc1a7b052a237a6653449a7c5f18edf19a3a79294c7d541c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/LICENSE"}], "id": "github:stanfordnlp/wge", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T11:18:16Z", "licenses": {"code": {"basis": "Stanford WGE source license; not a blanket license for all third-party tasks, embeddings or demonstration resources.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T11:23:05Z", "url": "https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/stanfordnlp/wge"}], "name": "Workflow-Guided Exploration research code", "notes": "The released repository includes training policies, experiment configuration and a MiniWoB sandbox extended for the 2018 paper. This is the WGE implementation, not another copy of the already cataloged Farama MiniWoB++ benchmark and not the original OpenAI MiniWoB release. Its documented dependency is PyTorch 0.1.12, and current runtime compatibility was not tested. Stanford code declares Apache-2.0; bundled third-party browser tasks and externally downloaded demonstration/embedding data retain their own terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/README.md", "target_id": "arxiv:1802.08802", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Research code that learns browser-control policies by constraining exploration with workflows induced from demonstrations, with a Selenium-based MiniWoB environment and experiment harness.", "topics": ["reinforcement-learning", "web-navigation", "workflow", "history"], "updated_at": null, "url": "https://github.com/stanfordnlp/wge"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "668949f121cb5a8ac262a8d3222a31eaee6692bfc0de9b86e60ae04f885ccc9b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/steel-dev/steel-browser/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/steel-dev/steel-browser/HEAD/LICENSE"}], "id": "github:steel-dev/steel-browser", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/steel-dev/steel-browser/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/steel-dev/steel-browser"}], "name": "Steel Browser", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/steel-dev/steel-browser/HEAD/README.md", "target_id": "organization:steel", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Open-source browser API managing Chrome sessions for agents, reachable through Puppeteer, Playwright or Selenium over CDP, with persisted cookies, proxy rotation, stealth and a session debugging UI.", "topics": ["sandbox", "web-navigation"], "updated_at": null, "url": "https://github.com/steel-dev/steel-browser"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "b92abb5fdf714e569cf9f0d2c85ac3fb91ed67365e371a4b322aafe6fab39a1d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/steel-dev/surf.new/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "efae611b1201878994c77b110867dcb65fb24a098c398bf5dec6138a47662bbc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/steel-dev/surf.new/HEAD/LICENSE"}], "id": "github:steel-dev/surf.new", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/steel-dev/surf.new/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/steel-dev/surf.new"}], "name": "Surf New", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Next.js and FastAPI playground from Steel for trying web agents in a controlled browser environment, with live views of browser interaction and support for providers such as Claude and GPT-4.", "topics": ["web-navigation"], "updated_at": null, "url": "https://github.com/steel-dev/surf.new"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "462e2b627989a969f7e859e22040a83d38e7c3eac12305d844563b0925f8a787", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/stepfun-ai/gelab-zero/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "08095ca875809a84572af7129caef5d4f4b052cc32ced7bdeb010fca125e07f0", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/stepfun-ai/gelab-zero/HEAD/LICENSE"}], "id": "github:stepfun-ai/gelab-zero", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/stepfun-ai/gelab-zero/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/stepfun-ai/gelab-zero"}], "name": "GELab Zero", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Open mobile GUI agent stack pairing a 4B model that runs on consumer hardware with local multi-device infrastructure, an MCP server, and AndroidDaily benchmark tasks.", "topics": ["mobile-navigation"], "updated_at": null, "url": "https://github.com/stepfun-ai/gelab-zero"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["Python code", "application API calls"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50", "population": "original AppWorld benchmark (2024 paper)", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2407.18901", "unit": "tasks", "value": 750}, {"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50", "population": "original AppWorld environment (2024 paper)", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2407.18901", "unit": "applications", "value": 9}], "evaluation": "State-based unit tests check task outcomes and unintended changes.", "evidence_urls": ["https://raw.githubusercontent.com/stonybrooknlp/appworld/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"], "focus": ["interactive coding", "cross-app tool use"], "notes": "This is an adjacent API/coding benchmark, not a screenshot-only task suite. Preserve the named release; task counts here refer to the original paper.", "observations": ["API documentation", "execution feedback"], "setting": "interactive"}, "leaderboard_url": "https://appworld.dev/appworld/leaderboard"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:45:03Z", "role": "primary", "sha256": "f4049141b6e520217395b7ccd9f27048c8e0f90331690dcbef300d2bc4ff7bdd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/stonybrooknlp/appworld/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/StonyBrookNLP/appworld/HEAD/LICENSE"}], "id": "github:stonybrooknlp/appworld", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:45:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/StonyBrookNLP/appworld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/StonyBrookNLP/appworld"}], "name": "AppWorld", "notes": "Primarily API and coding interaction; included as an adjacent comparison for GUI and hybrid agents.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/stonybrooknlp/appworld/HEAD/README.md", "target_id": "arxiv:2407.18901", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "Simulated applications, user data, and state-based task checks for evaluating agents that interact through APIs and code.", "topics": ["evaluation", "tool-integration", "cross-app"], "updated_at": null, "url": "https://github.com/StonyBrookNLP/appworld"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "cd42a4bd158fe3f0a1b0f376b9428d6a5fcb81b2", "repository_url": "https://github.com/sunblaze-ucb/AgentSynth"}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:12Z", "role": "primary", "sha256": "42e8b8dead7caea3f8b21e93a075e8c895dfc30a25e1470bf97a90940bbf0106", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:00Z", "role": "primary", "sha256": "e08f08a99b8cb3a4bb857a9fac976b2820f9ad406d66675b0286e880dfeeceaa", "supports": ["identity"], "url": "https://api.github.com/repos/sunblaze-ucb/AgentSynth"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:20Z", "role": "primary", "sha256": "b0998652a9b6b465444c2bbeef015bdbef50e83984a3bd15725f0e811645ca62", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/sunblaze-ucb/AgentSynth/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/LICENSE"}], "id": "github:sunblaze-ucb/agentsynth", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:12Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/sunblaze-ucb/AgentSynth"}], "name": "AgentSynth task-generation pipeline", "notes": "Released code includes task-generation and training-format conversion scripts. OSWorld/InSTA environments and model services require separate setup; the repository is not a bundled VM or model checkpoint. Code and the separately published dataset have distinct license declarations.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/README.md", "target_id": "arxiv:2506.14205", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Generates synthetic computer-use tasks and records trajectories in OSWorld desktop environments and InSTA browser environments.", "topics": ["data-generation", "long-horizon", "automation"], "updated_at": null, "url": "https://github.com/sunblaze-ucb/AgentSynth"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "9f851237932688c15d261c016d51ff5ae80c43c5d0b4eaccac0b48d941118d80", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Tencent/POINTS-GUI/HEAD/README.md"}], "id": "github:tencent/points-gui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Tencent/POINTS-GUI"}], "name": "POINTS-GUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Tencent/POINTS-GUI/HEAD/README.md", "target_id": "arxiv:2602.06391", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "POINTS-GUI-G-8B GUI grounding model, trained from a POINTS-1.5 base without native grounding through a data pipeline that unifies coordinates, filters noise and adds hard cases.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/Tencent/POINTS-GUI"}
{"aliases": ["github:mnotgod96/appagent", "mnotgod96/AppAgent", "https://github.com/mnotgod96/AppAgent"], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "ac94cca1f69807d2d31d30bd5521fbf1bda6155996881054df87a85522cd8f5b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/TencentQQGYLab/AppAgent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "d9e2b107cbaa3599336510dbb6703cb30210f1b4345221e76cf75cba85c2d113", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/TencentQQGYLab/AppAgent/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:13:57Z", "role": "primary", "sha256": "cd15c021e5023f43c17cc6365a58177c08ae91ac7e854c64c0e531f4ed3c5d5f", "supports": ["aliases"], "url": "https://api.github.com/repos/mnotgod96/AppAgent"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:13:58Z", "role": "primary", "sha256": "cd15c021e5023f43c17cc6365a58177c08ae91ac7e854c64c0e531f4ed3c5d5f", "supports": ["aliases"], "url": "https://api.github.com/repos/TencentQQGYLab/AppAgent"}], "id": "github:tencentqqgylab/appagent", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/TencentQQGYLab/AppAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/TencentQQGYLab/AppAgent"}], "name": "AppAgent", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/TencentQQGYLab/AppAgent/HEAD/README.md", "target_id": "arxiv:2312.13771", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Multimodal agent that operates Android apps through taps and swipes without back-end access, building a reusable knowledge base from autonomous exploration or from watching human demonstrations.", "topics": ["mobile-navigation", "memory"], "updated_at": null, "url": "https://github.com/TencentQQGYLab/AppAgent"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["web browsing", "code execution", "coworker communication"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/README.md", "population": "workspace images documented by the README", "qualifier": "exact", "unit": "task images", "value": 175}], "evaluation": "Task-specific evaluator scripts run within the prepared workplace environment.", "evidence_urls": ["https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/README.md"], "focus": ["professional workflows", "multi-tool tasks"], "notes": "Each task is packaged as a Docker image. Record task-image and server versions; the environment includes simulated coworkers and external application state.", "observations": [], "setting": "interactive"}, "leaderboard_url": "https://the-agent-company.com/#/leaderboard"}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:34Z", "role": "primary", "sha256": "d1646f67d49d38ef1e81bf7b5057dfc3d6f172b4c8a7413a2cc78741ccb08723", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "2d6976a9b482bb14d9cd4f0d97ce6e14305496f1649b8667cb91d027c670930f", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/LICENSE"}], "id": "github:theagentcompany/theagentcompany", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:34Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/TheAgentCompany/TheAgentCompany"}], "name": "TheAgentCompany", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/README.md", "target_id": "arxiv:2412.14161", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Self-contained workplace environment for evaluating agents on professional tasks involving websites, files, code, and simulated coworkers.", "topics": ["evaluation", "workflow", "cross-app"], "updated_at": null, "url": "https://github.com/TheAgentCompany/TheAgentCompany"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web", "android", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:55Z", "role": "primary", "sha256": "d69da19edaed185ad3071a97db3bb3fcfbdb14c0b514b7d27794b39fbe8e2541", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/thu-ml/MLA-Trust/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:46Z", "role": "primary", "sha256": "240fb6820e7bcb0753617bcad6ffbd4a9ba83ca98e1e6aaa06c7b197b244bf16", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/thu-ml/MLA-Trust/HEAD/LICENSE"}], "id": "github:thu-ml/mla-trust", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T10:59:55Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-18T11:03:46Z", "url": "https://raw.githubusercontent.com/thu-ml/MLA-Trust/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/thu-ml/MLA-Trust"}], "name": "MLA-Trust", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/thu-ml/MLA-Trust/HEAD/README.md", "target_id": "arxiv:2506.01616", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Unified evaluation framework measuring truthfulness, controllability, safety, and privacy of multimodal LLM agents across interactive web and mobile GUI tasks.", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": null, "url": "https://github.com/thu-ml/MLA-Trust"}
{"aliases": [], "authors": [], "details": {"leaderboard_url": "https://docs.google.com/spreadsheets/d/e/2PACX-1vRR3Wl7wsCgHpwUw1_eUXW_fptAPLL3FkhnW_rua0O1Ji_GIVrpTjY5LaKAhwO-WeARjnY_KNw0SYNJ/pubhtml"}, "environments": ["cross-platform", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "1b7d1f1cd7a74ac9497c639324ec342dd3f8769e9c0c4cd2cd92a4eb6740ac98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/THUDM/AgentBench/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/AgentBench/HEAD/LICENSE"}], "id": "github:thudm/agentbench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/THUDM/AgentBench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/AgentBench"}], "name": "AgentBench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/AgentBench/HEAD/README.md", "target_id": "arxiv:2308.03688", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "Agent evaluation suite with containerized tasks and a function-calling interface, including operating-system and shopping environments.", "topics": ["evaluation", "tool-integration"], "updated_at": null, "url": "https://github.com/THUDM/AgentBench"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["Android operations through the agent framework"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md", "population": "benchmark task collection", "qualifier": "exact", "unit": "tasks", "value": 138}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md", "population": "benchmark application set", "qualifier": "exact", "unit": "applications", "value": 9}], "evaluation": "The evaluation pipeline produces success, subgoal success, reversed redundancy, and reasonable-operation metrics; result generation requires a configured judge model.", "evidence_urls": ["https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md"], "focus": ["Android task completion", "subgoal completion", "operation efficiency"], "notes": "The benchmark uses predefined virtual-device images and apps chosen to work without network/login requirements. XML and SoM modes should be reported separately. These counts concern evaluation tasks, not the separately released Android Instruct training set.", "observations": ["UI XML", "set-of-marks views"], "setting": "interactive"}, "evaluation_type": "benchmark task evaluation", "leaderboard_url": "https://docs.google.com/spreadsheets/d/1Zv6mBfd4Ibt8mke24K6zAFrXe4AvEFaBLW3hpJqSgjw/edit?gid=0#gid=0", "metrics": ["task success"], "task_count": 138}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "340680a9a7833d103234b8367195deebd29bfec7e25ebc50c2dc04fc3627e55b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "340680a9a7833d103234b8367195deebd29bfec7e25ebc50c2dc04fc3627e55b", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:58Z", "role": "primary", "sha256": "1bbbee919161337e2456c587c6be809214406482cd9f1292a9eb62428f6961f3", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/LICENSE"}], "id": "github:thudm/android-lab", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:58Z", "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/Android-Lab"}], "name": "AndroidLab", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md", "target_id": "arxiv:2410.24024", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Reproducible Android benchmark of 138 tasks across nine apps on predefined virtual devices, run via AVD on Mac or Docker on Linux, released with the Android Instruct training set.", "topics": ["evaluation", "mobile-navigation"], "updated_at": null, "url": "https://github.com/THUDM/Android-Lab"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "9d18b958cfefd47100aa32ab14374c22ababb997a637a80da98b1977dec83fe0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/AutoWebGLM/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "0d97778d2d5e7439d1fbb0ce628565e7bd49886b2e86e6e6ca3d481c7fdc3256", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/AutoWebGLM/HEAD/LICENSE"}], "id": "github:thudm/autowebglm", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/THUDM/AutoWebGLM/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/AutoWebGLM"}], "name": "AutoWebGLM", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Web navigation agent built on ChatGLM3-6B that simplifies HTML for the model, is trained on hybrid human and AI data with RL and rejection sampling, and comes with the bilingual AutoWebBench.", "topics": ["web-navigation"], "updated_at": null, "url": "https://github.com/THUDM/AutoWebGLM"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f025e4bc9d920ec982525394c69b87043aae87fbe719de4b9e973408a44211e2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c6a77f53ba7933f07b95807e972e50259921344e777e52a6a8bc690373df2d74", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/LICENSE"}], "id": "github:thudm/cogagent", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/CogAgent"}], "name": "CogAgent", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/README.md", "target_id": "arxiv:2312.08914", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Visual language model for GUI understanding and interaction.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://github.com/THUDM/CogAgent"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "8b327dcd58c18e48e1e7f5b19ebf5673da4b5ff9320aa960e391844adf3a8a55", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/ComputerRL/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "a8d1d88a0f9cce3cdbff73b03a80dc5e5d69e3ba2443946c268815a550c9e0ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/ComputerRL/HEAD/LICENSE"}], "id": "github:thudm/computerrl", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/THUDM/ComputerRL/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/ComputerRL"}], "name": "ComputerRL", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/ComputerRL/HEAD/README.md", "target_id": "arxiv:2508.14040", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Reinforcement-learning infrastructure for computer-use agents.", "topics": ["reinforcement-learning"], "updated_at": null, "url": "https://github.com/THUDM/ComputerRL"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "752fad8ebd8b8eda03648329a6fc2585e5bcb2b96875b8aa2ce7d9f3fff78153", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "1bbbee919161337e2456c587c6be809214406482cd9f1292a9eb62428f6961f3", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/LICENSE"}], "id": "github:thudm/mobilerl", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/MobileRL"}], "name": "MobileRL", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/README.md", "target_id": "arxiv:2509.18119", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Online agentic RL framework for mobile GUI agents built around the difficulty-adaptive ADAGRPO algorithm, with Dockerized AndroidWorld and AndroidLab evaluation and a MobileRL-9B checkpoint.", "topics": ["reinforcement-learning"], "updated_at": null, "url": "https://github.com/THUDM/MobileRL"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "d4a68b00acc4becf2a25688520595babea9b694b56541e709f066413b5b80aea", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/SCALE-CUA/HEAD/README.md"}], "id": "github:thudm/scale-cua", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/SCALE-CUA"}], "name": "SCALE-CUA (THUDM)", "notes": "A distinct project from OpenGVLab/ScaleCUA, which documents arXiv:2509.15221. Similar names do not establish a shared project identity.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/SCALE-CUA/HEAD/README.md", "target_id": "arxiv:2607.11185", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Computer-use training framework with VeriGen task synthesis, adaptive rollout sampling, and visual-context segmentation.", "topics": ["reinforcement-learning", "data-generation", "efficiency"], "updated_at": null, "url": "https://github.com/THUDM/SCALE-CUA"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform", "web", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "21afbc4086441dee131fe4b84986870f9b280413dd2a38df6e99406f1fc45586", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/VisualAgentBench/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/VisualAgentBench/HEAD/LICENSE"}], "id": "github:thudm/visualagentbench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/THUDM/VisualAgentBench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/VisualAgentBench"}], "name": "VisualAgentBench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/VisualAgentBench/HEAD/README.md", "target_id": "arxiv:2408.06327", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "Evaluation framework spanning visual GUI, embodied, and design tasks, including mobile and WebArena-based environments.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/THUDM/VisualAgentBench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "eb217b96c54c3c2ccfd3a27255c6fd740df40c2d6743ae811858339bf7f740b0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/WebRL/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "42496cf0cff06a7f9e794f4313a10a670bc48df9909ec6d57763cb902bb229ea", "supports": ["identity"], "url": "https://api.github.com/repos/THUDM/WebRL"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "bb1683c6fb5ff658c89f0ebf0f8baf8026bb703840e174c9e46b223c85a16856", "supports": ["release_status"], "url": "https://api.github.com/repos/THUDM/WebRL/git/trees/main?recursive=1"}], "id": "github:thudm/webrl", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/WebRL"}], "name": "WebRL", "notes": "Training code and checkpoint links are released. No root code-license declaration was found in the inspected repository metadata; bundled dependency licenses do not establish a license for the whole project. Weights have separate Hub terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/WebRL/HEAD/README.md", "target_id": "arxiv:2411.02337", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training pipeline that evolves web tasks from prior attempts and uses an outcome reward model to refine browser policies.", "topics": ["reinforcement-learning", "web-navigation", "data-generation"], "updated_at": null, "url": "https://github.com/THUDM/WebRL"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["OSWorld/PyAutoGUI desktop interaction"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/README.md", "population": "reported benchmark total; not a recount of unique released task IDs", "qualifier": "exact", "unit": "tasks as reported by the paper/README", "value": 150}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misuse.json", "population": "released misuse task-ID list; before injection/goal expansion", "qualifier": "exact", "unit": "task IDs", "value": 50}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_injection.json", "population": "released injection task-ID list; before injection/goal expansion", "qualifier": "exact", "unit": "task IDs", "value": 10}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misbehavior.json", "population": "released misbehavior task-ID list; before injection/goal expansion", "qualifier": "exact", "unit": "task IDs", "value": 50}], "evaluation": "Automated LLM judge scores both task accuracy and safety from saved execution traces. Keep judge model/prompt version and harm category with a result; the scripts separately compute predicted metrics. This is an interactive desktop safety benchmark, not a single offline grounding score.", "evidence_urls": ["https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/README.md", "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/run.py", "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misuse.json", "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_injection.json", "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misbehavior.json"], "focus": ["deliberate misuse", "prompt-injection safety", "unintended harmful behavior during normal tasks"], "notes": "README/paper reports 150 tasks, whereas retained misuse/injection/misbehavior lists contain 50/10/50 IDs. run.py expands injection vectors/goals, so the released ID inventory is not a direct recount of paper settings; the complete expanded total was not reconstructed. Judge agreement F1 values in the README measure agreement with human labels, not agent success. Experiments use OSWorld desktop apps, with VMware documented as tested and screenshot/a11y/SOM observation choices. Code Apache-2.0 does not establish VM/app/data rights. Publisher asks that the benchmark not be used directly as agent training data.", "observations": ["screenshot", "accessibility tree", "screenshot plus accessibility tree", "set-of-marks"], "setting": "interactive"}, "evaluation_type": "automated judgment of accuracy and safety", "metrics": ["task completion", "safety"], "task_count": 150, "task_unit": "tasks reported by the paper/README; not a recount of released task IDs"}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T09:36:33Z", "role": "primary", "sha256": "d001158f58ac37b01c8470224db1567a52feb8010877305f7966637cf7400706", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:23:14Z", "role": "primary", "sha256": "86dd59a252fd654fc3244268a75de37b2c7f97e16f6b1de8c4cf885458a422f7", "supports": ["details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/run.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "8ff6092a51cc924c5225ace01c011c6481a2461eb0a47b7360eddaf6619b898c", "supports": ["details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misuse.json"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "b4c451236bb4ff36eb7e790dfa8cb517bffb245602bc441080314b13c8aba54f", "supports": ["details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_injection.json"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:05Z", "role": "primary", "sha256": "8eb6bade72696c3c0b53160a05b006a773c9e553ab5e401418b54d9acc98c0f2", "supports": ["details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misbehavior.json"}, {"method": "license_header", "retrieved_at": "2026-10-16T09:39:34Z", "role": "primary", "sha256": "7c3ab1eeeb073d92238ecb65ffaebe58ab5922929b115a7c0b64eb4f9a9afc2f", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/LICENSE"}], "id": "github:tml-epfl/os-harm", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T09:36:33Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T09:39:34Z", "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/tml-epfl/os-harm"}], "name": "OS-Harm", "notes": "The retained task-ID manifests have 50 misuse, 10 injection and 50 misbehavior IDs. Prompt-injection execution expands vectors/goals; the reported 150-task total must not be read as 150 unique released IDs.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/README.md", "target_id": "arxiv:2506.14866", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Benchmark for safety of computer-use agents across misuse, prompt injection, and unintended model behavior.", "topics": ["safety-security", "evaluation"], "updated_at": null, "url": "https://github.com/tml-epfl/os-harm"}
{"aliases": [], "authors": [], "details": {"availability": "Public training configurations, preprocessing and inference examples; full experiment-script release incomplete.", "license_review": {"code": {"evidence_urls": ["https://api.github.com/repos/TongUI-agent/TongUI-agent/git/trees/main?recursive=1"], "reason": "No project-level source-license declaration in the inspected tree; upstream dependency licenses do not establish this project’s grant."}}, "repository_revision": "82631b583180cdc870b4774d2495dd94f1558c46"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "f30a18d1a533b8f71f436dae2ee7369f8971422bd76527e069499c5bba911912", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "90c209a9758f7b527def072cfa78ebc87f1b49b436fad57c4c7bb5a6b66ae790", "supports": ["identity"], "url": "https://api.github.com/repos/TongUI-agent/TongUI-agent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "0c955a89dcda18652b09ce836c51c6ccaa007be6f6afeb48d21d01813cef1ec2", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/TongUI-agent/TongUI-agent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "b51acb70c95d079ccbe1de0a685ad7c8a4b3b5bed11e6162e6f7c2f60d78805e", "supports": ["summary", "details", "relations"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md"}], "id": "github:tongui-agent/tongui-agent", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/TongUI-agent/TongUI-agent"}], "name": "TongUI training and inference framework", "notes": "The public tree includes training configs, preprocessing, inference and serving examples. The project still marks release of all experiment/evaluation scripts as work in progress. This entry covers the implementation rather than counting its 3B/7B/32B model-size variants as separate new projects. The inspected project tree has no root source license; linked frameworks, data and checkpoints retain their own terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md", "target_id": "url:8c9aaafb918aaf8b4e92", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md", "target_id": "arxiv:2504.12679", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training configurations, preprocessing and inference examples for tutorial-trained GUI models, with a LLaMA-Factory fine-tuning workflow and screenshot-grounding interfaces.", "topics": ["grounding", "imitation-learning", "data-generation"], "updated_at": null, "url": "https://github.com/TongUI-agent/TongUI-agent"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/LICENSE", "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/README.md"], "reason": "The implementation and LICENSE live under MAI-UI/. A root-level missing license or the follow-up Qwen-UI-Agent documentation should not change the scoped declaration."}}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T05:29:14Z", "role": "primary", "sha256": "f644a5c0d0eb29b0ff8d8552dd505c5ecedf7c9cd9ab8cf2c617a1a14ecd0c14", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "fd90bc766d7e73889d14e8d3c5c92d384fa86e163222808f6e492509f00a3656", "supports": ["identity"], "url": "https://api.github.com/repos/Tongyi-MAI/MAI-UI"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:46Z", "role": "primary", "sha256": "4340b3d36bc56963fb8173079da8a0f7e91f195994d29fc47a9f7087cb56ffba", "supports": ["release_status"], "url": "https://api.github.com/repos/Tongyi-MAI/MAI-UI/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T05:29:15Z", "role": "primary", "sha256": "bf989559906a7926b77390082e4267ff7b2b494e8f599a9dc5640319eb45c2ce", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/LICENSE"}], "id": "github:tongyi-mai/mai-ui", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T05:29:14Z", "licenses": {"code": {"basis": "Retained MAI-UI/LICENSE and MAI-UI README; applies to the inspected implementation, subject to third-party components.", "declared": "Apache-2.0 (MAI-UI subdirectory)", "retrieved_at": "2026-10-17T05:29:15Z", "url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Tongyi-MAI/MAI-UI"}], "name": "MAI-UI inference and evaluation code", "notes": "The original implementation is in the MAI-UI/ subdirectory, with an Apache-2.0 LICENSE there. The current root README also promotes Qwen-UI-Agent, but that follow-up directory contains only documentation/display assets in the inspected tree. This entry does not claim the follow-up runtime, training system or weights are released.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/README.md", "target_id": "arxiv:2512.22047", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Published MAI-UI grounding and navigation-agent implementations with inference demos and evaluation scripts.", "topics": ["mobile-navigation", "grounding", "tool-integration"], "updated_at": null, "url": "https://github.com/Tongyi-MAI/MAI-UI"}
{"aliases": [], "authors": [], "details": {"availability": "Public task definitions, evaluator and container orchestration source.", "benchmark_protocol": {"actions": ["Android GUI operations", "MCP calls when enabled", "agent-user interaction when enabled"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md", "population": "README benchmark overview", "qualifier": "exact", "unit": "tasks", "value": 201}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md", "population": "README benchmark overview", "qualifier": "exact", "unit": "mobile applications", "value": 20}], "evaluation": "The publisher documents text-answer matching, backend database verification, local-storage inspection and application callbacks.", "evidence_urls": ["https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md"], "focus": ["Android task completion", "agent-user interaction", "hybrid GUI and MCP use"], "notes": "Publisher-reported inventory. GUI-only, agent-user and MCP-augmented configurations have different dependencies and task eligibility; preserve enabled modes and step budget with results.", "observations": ["Android application interface", "user-agent interaction when enabled"], "setting": "interactive"}, "leaderboard_url": "https://tongyi-mai.github.io/MobileWorld/#leaderboard", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/LICENSE"], "reason": "Repository code license; self-hosted applications, model APIs and MCP services keep their own terms."}}, "repository_revision": "e7eee8981135ecbe39005a231edf63574defd79c"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:16Z", "role": "primary", "sha256": "fbc306b3414dfdb230171358efd6eb8d67573fbb94e11487a4a7e971e2abf0d9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:15Z", "role": "primary", "sha256": "266e1fe5b5b1a4c1540ab0fd5eba206395084ee283f0079f424930a156cf8e18", "supports": ["identity"], "url": "https://api.github.com/repos/Tongyi-MAI/MobileWorld"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:08Z", "role": "primary", "sha256": "14500ed2dfa14bb5c34211090f0c8827e4a16c3233e35a2d695c0d7261f9358b", "supports": ["release_status"], "url": "https://api.github.com/repos/Tongyi-MAI/MobileWorld/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:09Z", "role": "primary", "sha256": "b09fcaf4fbc265223c2479c410beef2b98c5e687c78ecedf3452aa12ffc465f1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/LICENSE"}], "id": "github:tongyi-mai/mobileworld", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T09:12:16Z", "licenses": {"code": {"basis": "Repository code license; self-hosted applications, model APIs and MCP services keep their own terms.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T09:14:09Z", "url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Tongyi-MAI/MobileWorld"}], "name": "MobileWorld (Tongyi-MAI)", "notes": "This is Tongyi-MAI/MobileWorld, distinct from jacklishufan/MobileWorld (MobileWorldBench). Public source includes task definitions, verification logic and environment orchestration. The README reports 201 tasks across 20 applications; this is not an independent count of the evolving task tree. Linux/KVM or WSL2/KVM is required by the documented host setup, with macOS support in progress. MCP and simulated-user settings can require separate external services.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md", "target_id": "arxiv:2512.19432", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A containerized Android evaluation environment with GUI-only, agent-user interaction and MCP-augmented tasks, plus checks on application and backend state.", "topics": ["evaluation", "mobile-navigation", "human-in-the-loop", "tool-integration"], "updated_at": null, "url": "https://github.com/Tongyi-MAI/MobileWorld"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "12830b2eeb3edea7af9953b27987f3467fa90803c641e786db09a33643c714d7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/trycua/cua/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c0779290c1d4783169aa3dbfb55feb505e563ef8a004bbf55298ceffcfbda8d9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/trycua/cua/HEAD/LICENSE.md"}], "id": "github:trycua/cua", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/trycua/cua/HEAD/LICENSE.md"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/trycua/cua"}], "name": "Cua", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Monorepo behind Cua Spaces, which gives agents full macOS or Linux desktops, plus Cua Driver for operating native apps, Lume for Apple Silicon VMs, Cua Bench and the CUA-S1 decision models.", "topics": ["evaluation", "workflow"], "updated_at": null, "url": "https://github.com/trycua/cua"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "12830b2eeb3edea7af9953b27987f3467fa90803c641e786db09a33643c714d7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/trycua/lume/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "c0779290c1d4783169aa3dbfb55feb505e563ef8a004bbf55298ceffcfbda8d9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/trycua/lume/HEAD/LICENSE.md"}], "id": "github:trycua/lume", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/trycua/lume/HEAD/LICENSE.md"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/trycua/lume"}], "name": "Lume", "notes": "Lume is a component of the Cua repository. The legacy repository URL now serves the Cua monorepo README; counted as a component, not an independent project.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/trycua/lume/HEAD/README.md", "target_id": "github:trycua/cua", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "Virtual machine management for local computer-use environments.", "topics": ["sandbox"], "updated_at": null, "url": "https://github.com/trycua/lume"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "2c67c8b181e7196f3ca1e88888ffc5d6b32afca686a0e5c4353ad3f5af51ff9e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/tsinghua-fib-lab/FingerTip-20K/HEAD/README.md"}], "id": "github:tsinghua-fib-lab/fingertip-20k", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/tsinghua-fib-lab/FingerTip-20K"}], "name": "FingerTip-20K", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Over 20,000 human Android episodes across 500+ apps with user context and long-term usage patterns, plus tracks for proactive intent prediction and personalized execution.", "topics": ["data-generation"], "updated_at": null, "url": "https://github.com/tsinghua-fib-lab/FingerTip-20K"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "cfe9d8614b0b0558c37ba7c50371f6c47f119b914061ae815a9e3d796dd9ccfa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/UbiquitousLearning/DroidCall/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:01:50Z", "role": "primary", "sha256": "7a4ca3daa6d011ff2d24d993191882d2b5545d8476c7eddf396581c0fef77496", "supports": ["identity"], "url": "https://api.github.com/repos/UbiquitousLearning/DroidCall"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "d0717d706359969376889eac73ea8764b659f41a2b4f3c1a409110267a912526", "supports": ["release_status"], "url": "https://api.github.com/repos/UbiquitousLearning/DroidCall/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:16Z", "role": "primary", "sha256": "2832cc5333f7166924a2823aa38a286d918b39b7fef2ca7af6521766c68a5303", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UbiquitousLearning/DroidCall/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "08a6c5cfa655e1ecd774a41a9a815e845ad0afd47b289378bc1d6ee66c81dda9", "supports": ["relations"], "url": "https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:02:16Z", "role": "primary", "sha256": "2832cc5333f7166924a2823aa38a286d918b39b7fef2ca7af6521766c68a5303", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UbiquitousLearning/DroidCall/HEAD/LICENSE"}], "id": "github:ubiquitouslearning/droidcall", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T02:02:16Z", "url": "https://raw.githubusercontent.com/UbiquitousLearning/DroidCall/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/UbiquitousLearning/DroidCall"}], "name": "DroidCall generation and evaluation code", "notes": "Source utilities are published under MIT. This component operates through Android intent functions and is marked adjacent to visual GUI agents. Training data is separately released under its own card declaration; no checkpoint release is inferred from the fine-tuning instructions.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md", "target_id": "arxiv:2412.00402", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "Data-generation, fine-tuning and evaluation utilities for teaching small language models to invoke Android intents through predefined functions.", "topics": ["tool-integration", "data-generation", "evaluation"], "updated_at": null, "url": "https://github.com/UbiquitousLearning/DroidCall"}
{"aliases": ["github:eric-ai-lab/screen-point-and-read", "eric-ai-lab/Screen-Point-and-Read", "https://github.com/eric-ai-lab/Screen-Point-and-Read"], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "7567287ceee0ad2d3d5605abb44ac50463c1008f9b6309393695763cf2ce89bb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/eric-ai-lab/Screen-Point-and-Read/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "4748fb7d92f7a3a2a4a93353db96f6a2df547e5ea9ae87afa811c5423e96f363", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UCSB-AI/Screen-Point-and-Read/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:01Z", "role": "primary", "sha256": "868e18717ff76355c6552a5629b15de9fed1add1262fa5801ecabd7a7fb333d4", "supports": ["aliases"], "url": "https://api.github.com/repos/eric-ai-lab/Screen-Point-and-Read"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:02Z", "role": "primary", "sha256": "868e18717ff76355c6552a5629b15de9fed1add1262fa5801ecabd7a7fb333d4", "supports": ["aliases"], "url": "https://api.github.com/repos/UCSB-AI/Screen-Point-and-Read"}], "id": "github:ucsb-ai/screen-point-and-read", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/UCSB-AI/Screen-Point-and-Read/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/UCSB-AI/Screen-Point-and-Read"}], "name": "Screen Point-and-Read / Tree-of-Lens", "notes": "The canonical repository now belongs to UCSB-AI. Its README lists additional setup documentation and evaluation scripts as unfinished.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/eric-ai-lab/Screen-Point-and-Read/HEAD/README.md", "target_id": "arxiv:2406.19263", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screen-reading benchmark and Tree-of-Lens implementation for interpreting user-indicated regions and their layout context.", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": null, "url": "https://github.com/UCSB-AI/Screen-Point-and-Read"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "91cffd585c667d169486506a785c6cbc3068fe28409efc97abf8911f9d0bd5a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/UITron-hub/TreeCUA/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UITron-hub/TreeCUA/HEAD/LICENSE"}], "id": "github:uitron-hub/treecua", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/UITron-hub/TreeCUA/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/UITron-hub/TreeCUA"}], "name": "TreeCUA", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/UITron-hub/TreeCUA/HEAD/README.md", "target_id": "arxiv:2602.09662", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Framework for generating GUI trajectories through tree-structured exploration and verification.", "topics": ["data-generation"], "updated_at": null, "url": "https://github.com/UITron-hub/TreeCUA"}
{"aliases": [], "authors": [], "details": {"leaderboard_url": "https://unipat.ai/benchmarks/SaaS-Bench"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "8cfab6fca80090adb47c4ef634f89eedd49a6204dbdc2b5b8e0d6f04c32db289", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/UniPat-AI/SaaS-Bench/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "971246559df3f118e367cb299d3544f94e7c3d30db39f05c0c164589af4f3a88", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UniPat-AI/SaaS-Bench/HEAD/LICENSE"}], "id": "github:unipat-ai/saas-bench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/UniPat-AI/SaaS-Bench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/UniPat-AI/SaaS-Bench"}], "name": "SaaS-Bench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/UniPat-AI/SaaS-Bench/HEAD/README.md", "target_id": "arxiv:2605.15777", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "106 browser-driven business workflows across 23 self-hosted SaaS apps in six domains, scored by per-task verifiers that read each app's database and give partial credit.", "topics": ["evaluation", "workflow"], "updated_at": null, "url": "https://github.com/UniPat-AI/SaaS-Bench"}
{"aliases": [], "authors": [], "details": {"availability": "Published source; official release metadata inspected", "release_tag": "v0.39.1"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "4a9783ee8ebee10caaedfe2bdd2182539368e60e8eefadaee4cb7bab7fc75c12", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/vercel-labs/agent-browser/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "6b1120e122d5d110b0da568fbbf0560fd81a4cf8d50602c397a8382617343329", "supports": ["identity"], "url": "https://api.github.com/repos/vercel-labs/agent-browser"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "6a90f0b79089f895fab5d35443fd113faa056f404c807966b23210c1c427e62e", "supports": ["release_status"], "url": "https://api.github.com/repos/vercel-labs/agent-browser/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:55Z", "role": "primary", "sha256": "1a4213f48240031c7837e9651bda8cebb84d69d0d97a65f012082a53a7a1f702", "supports": ["release_status"], "url": "https://api.github.com/repos/vercel-labs/agent-browser/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:56Z", "role": "primary", "sha256": "014bb31e83d5c2e76aea1cc6e82217346ab41362f32cb355ad0f5c10aa0aeaff", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/vercel-labs/agent-browser/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:00:56Z", "role": "primary", "sha256": "014bb31e83d5c2e76aea1cc6e82217346ab41362f32cb355ad0f5c10aa0aeaff", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/vercel-labs/agent-browser/HEAD/LICENSE"}], "id": "github:vercel-labs/agent-browser", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T03:59:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T04:00:56Z", "url": "https://raw.githubusercontent.com/vercel-labs/agent-browser/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/vercel-labs/agent-browser"}], "name": "agent-browser", "notes": "The repository publishes a Rust-based CLI and platform-specific release binaries. The inspected release is v0.39.1. Apache-2.0 is declared for the code; Chrome and other external browser components retain separate terms. No package or browser was installed.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A browser-automation command-line interface that exposes browser state and interaction commands for AI agents.", "topics": ["web-navigation", "automation", "tool-integration"], "updated_at": null, "url": "https://github.com/vercel-labs/agent-browser"}
{"aliases": ["VeriGUI", "VeriWeb"], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "e9a0866c8d9fa1759dedebd5ea3a4176009e51eb669f5877540524e53ac1077d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/VeriGUI-Team/VeriGUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/VeriGUI-Team/VeriGUI/HEAD/LICENSE"}], "id": "github:verigui-team/verigui", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/VeriGUI-Team/VeriGUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/VeriGUI-Team/VeriGUI"}], "name": "VeriWeb (formerly VeriGUI)", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Verifiable web information-seeking benchmark renamed from VeriGUI; desktop evaluation is described as future work.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/VeriGUI-Team/VeriGUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:55Z", "role": "primary", "sha256": "2403ef3ec300e257dc631e3c569fc9aabaab1d3f385ac37abe1e4021aaae6e21", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/VeriSafeAgent/VeriSafeAgent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T13:06:20.028238+00:00", "role": "primary", "sha256": "9b04efe4136a39e3e5f21a1d3bb8a44075f1370bdbeec6a4307783f3836f8311", "supports": ["summary", "relations", "notes"], "url": "https://arxiv.org/html/2503.18492v2"}], "id": "github:verisafeagent/verisafeagent", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T10:59:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/VeriSafeAgent/VeriSafeAgent"}], "name": "VeriSafe Agent", "notes": "The verification method requires developers to supply app-state predicates and transition hooks. The paper-to-repository relation is supported by the exact-version paper, not inferred from the repository name.", "published_at": null, "relations": [{"evidence_url": "https://arxiv.org/html/2503.18492v2", "target_id": "arxiv:2503.18492", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Runtime verification framework for Android GUI agents that checks proposed actions against logical rules derived from user instructions and developer-defined app states.", "topics": ["safety-security", "mobile-navigation", "state-tracking"], "updated_at": null, "url": "https://github.com/VeriSafeAgent/VeriSafeAgent"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "262947081ddd17dbf5bb3d072574a19bbf30510c42c3ff42c9851a748fd37546", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/VisualWebBench/VisualWebBench/HEAD/README.md"}], "id": "github:visualwebbench/visualwebbench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/VisualWebBench/VisualWebBench"}], "name": "VisualWebBench", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/VisualWebBench/VisualWebBench/HEAD/README.md", "target_id": "arxiv:2404.05955", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Evaluation of multimodal models on visual web understanding tasks.", "topics": ["evaluation", "screen-understanding"], "updated_at": null, "url": "https://github.com/VisualWebBench/VisualWebBench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "5f460f9b2a18315d07b8c48d78dcc764892d09d9250e95f5e591b5966fe59328", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/vivo-ai-lab/BlueLM-GUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "6036aa17e52be8051fc3c8ede37fc4d0b2bc2589518be7623566c75868612178", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/vivo-ai-lab/BlueLM-GUI/HEAD/LICENSE"}], "id": "github:vivo-ai-lab/bluelm-gui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/vivo-ai-lab/BlueLM-GUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/vivo-ai-lab/BlueLM-GUI"}], "name": "BlueLM-GUI", "notes": "The repository lists the agent as in-house. Public model weights were not established by this review; linked benchmark assets have their own access terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/vivo-ai-lab/BlueLM-GUI/HEAD/README.md", "target_id": "arxiv:2609.12394", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile GUI agent research release with technical report, benchmark data, and evaluation tooling from vivo AI Lab.", "topics": ["mobile-navigation", "reinforcement-learning"], "updated_at": null, "url": "https://github.com/vivo-ai-lab/BlueLM-GUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "4f578ecd5cbba12797586892a9190e480291bd267a180f0ce56847764eaac524", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/readme.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "d1a708ddcd4afc4e40a5d66f2d56835a12d815515ab6a61578e541ed2c0483bb", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/LICENSE"}], "id": "github:vivo/dimo-gui", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/vivo/DiMo-GUI"}], "name": "DiMo-GUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/readme.md", "target_id": "arxiv:2507.00008", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI grounding research using modality-aware visual reasoning and test-time scaling.", "topics": ["grounding", "efficiency"], "updated_at": null, "url": "https://github.com/vivo/DiMo-GUI"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["browser element actions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md", "population": "full VWA suite", "qualifier": "exact", "unit": "tasks", "value": 910}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md", "population": "one task from each template type", "qualifier": "exact", "unit": "human trajectories", "value": 233}], "evaluation": "Execution-based completion checks for visually grounded web tasks.", "evidence_urls": ["https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md"], "focus": ["multimodal web task completion"], "notes": "The 233 demonstration trajectories are not the full 910-task evaluation set. Observation modes and captioning components differ across baseline agents.", "observations": ["screenshots with set-of-mark annotations", "caption-augmented accessibility tree"], "setting": "interactive"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "182529c2fce1e41f3a2162727205c54da8b3d2381b2273d726170ee9d8cf65cc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md"}], "id": "github:web-arena-x/visualwebarena", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/web-arena-x/visualwebarena"}], "name": "VisualWebArena", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md", "target_id": "arxiv:2401.13649", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Visually grounded web tasks extending the WebArena evaluation setting.", "topics": ["evaluation", "screen-understanding"], "updated_at": null, "url": "https://github.com/web-arena-x/visualwebarena"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["browser element actions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/README.md", "population": "canonical WebArena evaluation suite", "qualifier": "exact", "unit": "task examples", "value": 812}], "evaluation": "Functional correctness of completed web tasks.", "evidence_urls": ["https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"], "focus": ["web task completion"], "notes": "The canonical repository documents self-hosted websites and environment resets. Its README recommends BrowserGym/AgentLab for newer infrastructure; harness changes must be recorded.", "observations": ["HTML or accessibility tree"], "setting": "interactive"}, "leaderboard_url": "https://docs.google.com/spreadsheets/d/1M801lEpBbKSNwP-vDBkC_pF7LdyGU1f_ufZb_NWNBZQ/edit?usp=sharing"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "081c826497f9b7da4952771ee128c2215ec3532379db0fe9f7fcbd899d9c9287", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/LICENSE"}], "id": "github:web-arena-x/webarena", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/web-arena-x/webarena"}], "name": "WebArena", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/README.md", "target_id": "arxiv:2307.13854", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Self-hosted web environments and tasks for evaluating autonomous web agents.", "topics": ["evaluation", "web-navigation"], "updated_at": null, "url": "https://github.com/web-arena-x/webarena"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "8cd83b12f4e1c1b22c27b6ae3ee5415564e61ff4072349ec512982a548ea6ec3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/web-infra-dev/midscene/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "9d26ef3ec511840e10eeaaf1e1b211e196c838823baf6bf9033761f27ab349ce", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/web-infra-dev/midscene/HEAD/LICENSE"}], "id": "github:web-infra-dev/midscene", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/web-infra-dev/midscene/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/web-infra-dev/midscene"}], "name": "Midscene.js", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Visual automation toolkit for web, mobile, and desktop interfaces.", "topics": ["automation", "grounding"], "updated_at": null, "url": "https://github.com/web-infra-dev/midscene"}
{"aliases": [], "authors": [], "details": {"availability": "Public WebVLN training/evaluation and environment-wrapper source; external data/models unverified.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/WebVLN/WebVLN/HEAD/README.md", "https://api.github.com/repos/WebVLN/WebVLN/git/trees/main?recursive=1"], "reason": "README MIT badge is not backed by a project license file or component-scoped statement in the inspected release."}}, "repository_revision": "59aba284f5888fb7730e1ff07a7e682a1bd1968a"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "cb98feab09648856ab30d8eb739a29e837673cc5b4a6642b7ba3a8943ad71053", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/WebVLN/WebVLN/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "112e67cc9f865b9640da3827e32eaba5a50f2feb8d85c20018dccbcafa6fd3e5", "supports": ["identity"], "url": "https://api.github.com/repos/WebVLN/WebVLN"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "966adf983bae6052c271f7153c30776a29f7272cf55d05e0175713f02d1900fb", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/WebVLN/WebVLN/git/trees/main?recursive=1"}], "id": "github:webvln/webvln", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/WebVLN/WebVLN"}], "name": "WebVLN research implementation", "notes": "The released tree contains navigation models, an environment wrapper, scoring utilities and train/eval scripts. Full WebVLN-v1 data and pretrained files are linked through an external Drive folder; their complete contents and data/weight terms were not verified, so no separate benchmark download or checkpoint is approved here. The README MIT badge has no matching repository license file or explicit component scope; code permissions remain unestablished.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/WebVLN/WebVLN/HEAD/README.md", "target_id": "arxiv:2312.15820", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training and evaluation code for question-guided website navigation that combines visual observations with web-specific HTML information.", "topics": ["web-navigation", "grounding", "evaluation"], "updated_at": null, "url": "https://github.com/WebVLN/WebVLN"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:22Z", "role": "primary", "sha256": "813b7580cdae140917db573522b83a7986448025b31064442d34757292a6935c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/weizhepei/WebAgent-R1/HEAD/WebAgent-R1/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:19Z", "role": "primary", "sha256": "db1f71491c1d9482f403db30ea5599bd76e93d1db02c4ba8134d1a5cae5f0487", "supports": ["identity"], "url": "https://api.github.com/repos/weizhepei/WebAgent-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:39Z", "role": "primary", "sha256": "cc1a74037c4a2c4c094b11f6874ef67f88b24d14293ec09568ac6f378035c0b0", "supports": ["release_status"], "url": "https://api.github.com/repos/weizhepei/WebAgent-R1/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:22Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/weizhepei/WebAgent-R1/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:05:22Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/weizhepei/WebAgent-R1/HEAD/LICENSE"}], "id": "github:weizhepei/webagent-r1", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:05:22Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T02:05:22Z", "url": "https://raw.githubusercontent.com/weizhepei/WebAgent-R1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/weizhepei/WebAgent-R1"}], "name": "WebAgent-R1", "notes": "The implementation is inside the WebAgent-R1 subdirectory and includes training, evaluation and WebArena setup components. Root source licensing is Apache-2.0. No downloadable author checkpoint was established by the reviewed README, so this entry represents code.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/weizhepei/WebAgent-R1/HEAD/WebAgent-R1/README.md", "target_id": "arxiv:2505.16421", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Environment setup, multi-turn reinforcement-learning and evaluation tooling for browser agents interacting with WebArena.", "topics": ["reinforcement-learning", "web-navigation", "evaluation"], "updated_at": null, "url": "https://github.com/weizhepei/WebAgent-R1"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "085f1a1810805bcade9555b1e5c374a7db1b8b1377afeeea285d927853b02f3d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Westlake-AGI-Lab/AppAgentX/HEAD/README.md"}], "id": "github:westlake-agi-lab/appagentx", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Westlake-AGI-Lab/AppAgentX"}], "name": "AppAgentX", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Westlake-AGI-Lab/AppAgentX/HEAD/README.md", "target_id": "arxiv:2503.02268", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile agent research using accumulated interaction experience.", "topics": ["mobile-navigation", "memory"], "updated_at": null, "url": "https://github.com/Westlake-AGI-Lab/AppAgentX"}
{"aliases": [], "authors": [], "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/LICENSE", "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/NOTICE"], "reason": "Record only the publisher's top-level Apache-2.0 declaration for Orbit-original code. NOTICE explicitly preserves per-component terms; neither bundled or external data nor model weights inherit this declaration. Its account of BrowserART licensing is not an independent resolution of upstream conflicting declarations."}}}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:41:22.467183+00:00", "role": "primary", "sha256": "b6743e3e9e5256d7afa8bb411baf246b60ba71bcfb84eda5a816fba3ddd7d78b", "supports": ["name", "kind", "summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T13:39:57.052202+00:00", "role": "primary", "sha256": "0436414b9bcc4307f3bcde476a544da6fe1dd5baa90939eed392942490fbfbf7", "supports": ["id", "url"], "url": "https://api.github.com/repos/wlanderson0/orbit"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:42:37.098639+00:00", "role": "primary", "sha256": "c47d046b129999da5e63b7723a3f3e2e0cc31c27ce74b45f71f8c6d0df1fc0d8", "supports": ["summary", "environments", "scope", "notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/orbit/scenarios/browser/browserart/tools.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T14:10:32.910533Z", "role": "primary", "sha256": "2cce76b22252a415d64c356b2062771e528b554dd83b6f88fca769f48fa8a826", "supports": ["summary", "environments", "scope", "notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/orbit/scenarios/desktop/osworld/tools.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T14:09:35.060985Z", "role": "primary", "sha256": "66515d7ee87eab5a2f42c26132d0bf3e2a3b9c006156d2378181cf91966a2903", "supports": ["notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/docs/benchmark-qualification.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T14:10:29.541892Z", "role": "primary", "sha256": "94c781028dc80b2a748d759fae2456f6a986da7487651a6bcbb179e8ad4fe023", "supports": ["notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/orbit/scenarios/browser/browserart/scorer.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:41:29.036076+00:00", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T14:09:31.550310Z", "role": "primary", "sha256": "f68a64f775b1f329416f2fe294bb07b5e475893f77d1cdb5300bdb08acf95410", "supports": ["licenses.code", "licenses.data", "notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/NOTICE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:28.833839+00:00", "role": "primary", "sha256": "3ea667f4187ba601011e5ff00c4d63676b6d9903d12b80a0d7d5dd9920f00c65", "supports": ["relations"], "url": "https://arxiv.org/html/2609.33102v1"}], "id": "github:wlanderson0/orbit", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-20T13:41:22.467183+00:00", "licenses": {"code": {"basis": "Complete root Apache License 2.0 text for Orbit-original code. NOTICE preserves separate third-party component terms, including MIT-vendored modules; this is not a uniform license for the entire dependency/data bundle.", "declared": "Apache-2.0", "retrieved_at": "2026-10-20T13:41:29.036076+00:00", "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/wlanderson0/orbit"}], "name": "ORBIT framework", "notes": "Reviewed version 1.0.3 source at commit 588b3035, not an asserted reproduction of the paper. BrowserART tools return accessibility trees, including the screenshot-named refresh. Qualification is scenario-specific: the project documents browser-judge and desktop-scoring limits; its opt-in exact-send oracle covers one derived local task, not SMTP delivery. Runtime profiles and scoring policies are separate. Apache-2.0 is the top-level Orbit code declaration; third-party components and external data/environments retain separate terms. No installation, release archive or live evaluation was tested.", "published_at": null, "relations": [{"evidence_url": "https://arxiv.org/html/2609.33102v1", "target_id": "arxiv:2609.33102", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Inspect-based multi-agent evaluation framework with role, scheduling and communication controls, BrowserART accessibility-tree actions, and desktop mouse, keyboard and screenshot tools.", "topics": ["evaluation", "safety-security", "web-navigation", "observability"], "updated_at": null, "url": "https://github.com/wlanderson0/orbit"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["efficiency scores"], "counts": [], "evaluation": "The scorer consumes OSWorld results and a specified maximum step budget; the README distinguishes single-action and grouped-action WES+ and WES-.", "evidence_urls": ["https://raw.githubusercontent.com/WukLab/osworld-human/HEAD/README.md"], "focus": ["action efficiency relative to humans"], "notes": "Reference solutions are published for evaluation and expressly discouraged for training. No task count is inferred from file count.", "observations": ["OSWorld result trajectories", "human reference actions"], "setting": "offline"}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "35c2f4dbed9d9bfc57f4c5a3f2fd70d313d6b463403b482c2fd4066d27b02cc1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/WukLab/osworld-human/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "21c86601da6e77a3326fbcd143e3c5d9a033cd59e6e13973250966dff6f51ae9", "supports": ["identity"], "url": "https://api.github.com/repos/WukLab/osworld-human"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:55Z", "role": "primary", "sha256": "c7b958d7a075adf5d282be08b68ad2b090568da342d4cb4b501ae808e54d0053", "supports": ["release_status"], "url": "https://api.github.com/repos/WukLab/osworld-human/git/trees/main?recursive=1"}], "id": "github:wuklab/osworld-human", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/WukLab/osworld-human"}], "name": "OSWorld-Human", "notes": "The repository publishes per-application reference JSON files and score.py. It explicitly states that these evaluation solutions should not be used for training. No root code or data-license declaration was found; no license is inherited from OSWorld, and the score script was not executed.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/WukLab/osworld-human/HEAD/README.md", "target_id": "arxiv:2506.16042", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Human reference trajectories and scoring code for measuring the action efficiency of computer-use agents on OSWorld tasks.", "topics": ["evaluation", "efficiency", "workflow"], "updated_at": null, "url": "https://github.com/WukLab/osworld-human"}
{"aliases": [], "authors": [], "details": {"availability": "Public synthesis and evaluation source; model checkpoints not verified.", "repository_revision": "1ec25af942a46ce46f6c00d596332e308198342f"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:20Z", "role": "primary", "sha256": "9a46741bbde1373ed4b87887cdadde98e05ea8795d34efe39fc25d69b4d50213", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:20Z", "role": "primary", "sha256": "ae1d5687a72a39eacbf4fe281b5048914e98adf7cc1517481a5acc8a56a81446", "supports": ["identity"], "url": "https://api.github.com/repos/Wuzheng02/GUI-CIDER"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:59Z", "role": "primary", "sha256": "355128b62a281074eebc14a1e3b85a260f06c2bf3c1a6ed140cba3b9aefaa76c", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/Wuzheng02/GUI-CIDER/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:24:47Z", "role": "primary", "sha256": "bffca456e649a08239f6a1a3be10a7d7c21bcb427e783ba2c304e4b8410d385a", "supports": ["environments", "summary", "notes"], "url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/data/dataset.md"}], "id": "github:wuzheng02/gui-cider", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T11:18:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Wuzheng02/GUI-CIDER"}], "name": "GUI-CIDER research code", "notes": "The source tree contains static/causal knowledge synthesis and action/knowledge evaluation scripts. Training procedures are delegated to LLaMA-Factory; no trained checkpoint is verified here. The README displays an MIT badge, but the inspected tree has no license file and no explicit component scope, so the code license remains unrecorded rather than assuming a blanket grant. Evaluation requires separately acquired original datasets and model paths.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md", "target_id": "arxiv:2605.28534", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Released data-synthesis and evaluation scripts that distill planning and causal GUI knowledge into text for mid-training mobile interface agents.", "topics": ["data-generation", "imitation-learning", "planning"], "updated_at": null, "url": "https://github.com/Wuzheng02/GUI-CIDER"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:14:29Z", "role": "primary", "sha256": "15d40c50df8493575a94aa2aa4913c654fea7b7f821d99b7bf443a64f1cc50d3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Wuzheng02/OS-Kairos/HEAD/README.md"}], "id": "github:wuzheng02/os-kairos", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T13:14:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Wuzheng02/OS-Kairos"}], "name": "OS-Kairos", "notes": "The documented evaluation modes distinguish model-simulated human feedback from actual human interaction; preserve that distinction when interpreting results.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Wuzheng02/OS-Kairos/HEAD/README.md", "target_id": "arxiv:2503.16465", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Android GUI-agent research implementation that uses confidence-guided interaction to choose between autonomous actions and human assistance.", "topics": ["human-in-the-loop", "safety-security", "mobile-navigation"], "updated_at": null, "url": "https://github.com/Wuzheng02/OS-Kairos"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:55Z", "role": "primary", "sha256": "282883d02391a86c5a4d0cdafdc1d33f10a5073f31eba8c852095601d0aeb939", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "notes"], "url": "https://raw.githubusercontent.com/Wuzheng02/VeriOS/HEAD/README.md"}], "id": "github:wuzheng02/verios", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T10:59:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Wuzheng02/VeriOS"}], "name": "VeriOS", "notes": "The released scripts evaluate recorded task inputs; the retained README does not establish a live Android execution environment. No runtime behavior or reported results were independently reproduced.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Wuzheng02/VeriOS/HEAD/README.md", "target_id": "arxiv:2509.07553", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Research code for proactive human–GUI-agent verification, including offline interaction-loop and dual-agent evaluation scripts plus LLaMA-Factory fine-tuning configurations.", "topics": ["safety-security", "human-in-the-loop", "mobile-navigation"], "updated_at": null, "url": "https://github.com/Wuzheng02/VeriOS"}
{"aliases": [], "authors": [], "details": {"availability": "Public federated-training/evaluation code and ungated annotation partitions; original images are separate dependencies.", "benchmark_protocol": {"actions": ["model predictions on mobile-agent examples"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/wwh0411/FedMABench/raw/main/README.md", "population": "released FedMABench partition families", "qualifier": "exact", "unit": "dataset families", "value": 6}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md", "population": "publisher benchmark overview", "qualifier": "exact", "unit": "heterogeneity types", "value": 3}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md", "population": "publisher benchmark overview", "qualifier": "exact", "unit": "federated learning algorithms", "value": 8}], "evaluation": "The provided evaluation workflow performs model inference over selected partitions and invokes test_swift_fed.py for accuracy scoring.", "evidence_urls": ["https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md", "https://huggingface.co/datasets/wwh0411/FedMABench/raw/main/README.md"], "focus": ["mobile-agent federated training", "data heterogeneity"], "notes": "The dataset-family count does not count tasks, apps, screenshots or client instances. Code and data licenses are unknown; external image sources are required.", "observations": ["mobile-agent examples with externally sourced screenshots"], "setting": "offline"}, "repository_revision": "417d8b0006830f423c7d07c213fc57277d75d966"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:17Z", "role": "primary", "sha256": "b49e2b419918078ef9c68f3c9becc6c13d8fe530c165a36bdd61e370746eeea8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:17Z", "role": "primary", "sha256": "3e1b2b643a263af129b3a4092a84efe47d3643f00509f7b9517da21a4aac0371", "supports": ["identity"], "url": "https://api.github.com/repos/wwh0411/FedMABench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:54Z", "role": "primary", "sha256": "4fb977a469cee3d5083da0b5bf2984fedc6b867146a90b20963985e08f9e9716", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/wwh0411/FedMABench/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:06Z", "role": "primary", "sha256": "df5379f4a9b350dad0c3d374d2f9a1d6eacd974b6d54750e2a14ecec1faec6e5", "supports": ["identity", "release_status"], "url": "https://huggingface.co/api/datasets/wwh0411/FedMABench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:06Z", "role": "primary", "sha256": "2874269995d8c9f5dabe266433ecf724bf54c27af599d899f3f0090b8ee37cdc", "supports": ["summary", "details.benchmark_protocol", "licenses.data"], "url": "https://huggingface.co/datasets/wwh0411/FedMABench/raw/main/README.md"}], "id": "github:wwh0411/fedmabench", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T11:18:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/wwh0411/FedMABench"}], "name": "FedMABench", "notes": "The public Hub lists six dataset families of JSON/JSONL partition annotations. The repository explicitly says unpacked screenshots are not provided and must be reconstructed from original AndroidControl and AITW sources. This is offline training/evaluation infrastructure rather than a live Android task environment. No repository license file or dataset-license declaration was found in the inspected sources; neither is inferred from the underlying ms-swift project.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md", "target_id": "arxiv:2503.05143", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A federated mobile-agent training and evaluation suite that varies user-data heterogeneity and application distribution, with released partition annotations and inference/scoring scripts.", "topics": ["evaluation", "imitation-learning", "mobile-navigation"], "updated_at": null, "url": "https://github.com/wwh0411/FedMABench"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "2cea1be73ea75acd46990ab1fb567fbd83ad58a2bc69ca858b207660df9fe8c1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/X-LANCE/META-GUI-baseline/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "01fb614f0536ea76c1c25809cc8cde66cde21a96ac6ca0ea81b2c093e2157d2a", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/X-LANCE/META-GUI-baseline/HEAD/LICENSE"}], "id": "github:x-lance/meta-gui-baseline", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/X-LANCE/META-GUI-baseline/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/X-LANCE/META-GUI-baseline"}], "name": "META-GUI baseline", "notes": "Its README describes mobile GUI agents without naming an operating system.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/X-LANCE/META-GUI-baseline/HEAD/README.md", "target_id": "arxiv:2205.11029", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training and evaluation pipeline for conversational mobile GUI agents, with action prediction, response generation, and dataset processing.", "topics": ["mobile-navigation", "human-in-the-loop"], "updated_at": null, "url": "https://github.com/X-LANCE/META-GUI-baseline"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "fa70fd96c57a163404b77431f89dd16864ed104d7160e347742ff87587229961", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/X-LANCE/Mobile-Env/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "58d1e17ffe5109a7ae296caafcadfdbe6a7d176f0bc4ab01e12a689b0499d8bd", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/X-LANCE/Mobile-Env/HEAD/LICENSE"}], "id": "github:x-lance/mobile-env", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/X-LANCE/Mobile-Env/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/X-LANCE/Mobile-Env"}], "name": "Mobile Env", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/X-LANCE/Mobile-Env/HEAD/README.md", "target_id": "arxiv:2305.08144", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "AndroidEnv-based platform for building GUI agent benchmarks and training agents, where task definition files emit instructions, rewards and episode ends from screen text, view hierarchy or log events.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/X-LANCE/Mobile-Env"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "ca402ffd3b05f72a6d72f47de52c962f92e410749a77939044f3301d4330af16", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "961819a89dacd1f4f161e1d2ebf616b538bf095a2eecfa3ef40b7ccec633fac4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}], "id": "github:x-plug/mobileagent", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/X-PLUG/MobileAgent"}], "name": "Mobile-Agent", "notes": "This repository contains multiple Mobile-Agent generations. Platform coverage differs by version; use the corresponding paper and implementation instructions.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/README.md", "target_id": "arxiv:2401.16158", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/README.md", "target_id": "arxiv:2406.01014", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/README.md", "target_id": "arxiv:2508.15144", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/README.md", "target_id": "arxiv:2602.16855", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile device agent family using visual interface interaction.", "topics": ["mobile-navigation", "planning"], "updated_at": null, "url": "https://github.com/X-PLUG/MobileAgent"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "55de0ed6627a5812d1f26d2a920e0af47afb17f8a99cae41b70a85ba30a23370", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/xbmxb/CoCo-Agent/HEAD/README.md"}], "id": "github:xbmxb/coco-agent", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xbmxb/CoCo-Agent"}], "name": "CoCo-Agent", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/xbmxb/CoCo-Agent/HEAD/README.md", "target_id": "arxiv:2402.11941", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Research implementation combining screen and layout perception with conditional action prediction for smartphone automation.", "topics": ["mobile-navigation", "screen-understanding"], "updated_at": null, "url": "https://github.com/xbmxb/CoCo-Agent"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "17fa475b96bdd00270c656072d3f8fc55d8433f0e3a1c2762110160ec21a5c35", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/xlang-ai/AgentTrek/HEAD/README.md"}], "id": "github:xlang-ai/agenttrek", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/AgentTrek"}], "name": "AgentTrek", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/AgentTrek/HEAD/README.md", "target_id": "arxiv:2412.09605", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Synthesizes browser-agent trajectories by replaying web tutorials as guidance, releasing the trajectory data with multimodal grounding and reasoning and an AgentTrek-32B agent.", "topics": ["data-generation", "video"], "updated_at": null, "url": "https://github.com/xlang-ai/AgentTrek"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dd9f36b818c1b7b6bd772ba8575f6c0ebab6d9e0d307edcc2f196f3e4ba502e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md"}], "id": "github:xlang-ai/aguvis", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/aguvis"}], "name": "Aguvis", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md", "target_id": "arxiv:2412.04454", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Vision-based GUI agent models with unified action representations.", "topics": ["grounding", "planning"], "updated_at": null, "url": "https://github.com/xlang-ai/aguvis"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "91fed11b234775bde91bb766ee65e1015897ced6668de8e80418fcee09f8f75d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenCUA/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "a53fdf6cdedc93c99879b2881d8626d61b09e9a03841b3dc6ebe4e267f4fb2bf", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenCUA/HEAD/LICENSE"}], "id": "github:xlang-ai/opencua", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/xlang-ai/OpenCUA/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/OpenCUA"}], "name": "OpenCUA", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/OpenCUA/HEAD/README.md", "target_id": "arxiv:2508.09123", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Open computer-use research stack including models and training resources.", "topics": ["data-generation", "planning"], "updated_at": null, "url": "https://github.com/xlang-ai/OpenCUA"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["mouse and keyboard via PyAutoGUI"], "counts": [{"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_all.json", "population": "test_all.json", "qualifier": "exact", "unit": "task definitions", "value": 369}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_nogdrive.json", "population": "test_nogdrive.json", "qualifier": "exact", "unit": "task definitions", "value": 361}], "evaluation": "Task-specific execution-based evaluators inspect application or system state.", "evidence_urls": ["https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50", "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_all.json", "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_nogdrive.json"], "focus": ["task completion", "cross-application workflows"], "notes": "Counts were computed from unique task identifiers in the two inspected manifests. The subsets differ; record the manifest, repository revision, agent interface, and step budget with a score.", "observations": ["screenshots"], "setting": "interactive"}, "leaderboard_url": "https://os-world.github.io/"}, "environments": ["desktop", "linux", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "83b8684c2f4ac4f31996b8b4752195fcb6cc5429f89d92398177b9b721d787fc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:34:59Z", "role": "primary", "sha256": "9ebc5187cbd727ef26c24626820076b102fff812863c640a67c467fea9542ab5", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_all.json"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:34:59Z", "role": "primary", "sha256": "fcb9497e93a8986407345d3012c872c9c2fed253420730fbea64ccfcace67dbb", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_nogdrive.json"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "a8d1d88a0f9cce3cdbff73b03a80dc5e5d69e3ba2443946c268815a550c9e0ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/LICENSE"}], "id": "github:xlang-ai/osworld", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/OSWorld"}], "name": "OSWorld", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/README.md", "target_id": "arxiv:2404.07972", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Execution-based evaluation of agents performing tasks in real desktop environments.", "topics": ["evaluation", "cross-app"], "updated_at": null, "url": "https://github.com/xlang-ai/OSWorld"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["click point or box center", "refusal represented by negative coordinates in the evaluator"], "counts": [], "evaluation": "The inspected Jedi runner maps predictions from resized-image coordinates back to the original image, then checks the predicted center against rectangle/polygon targets or the refusal rule. It reports overall and classification-group accuracy. This is offline grounding, not OSWorld task-execution success.", "evidence_urls": ["https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/README.md", "https://api.github.com/repos/xlang-ai/OSWorld-G/git/trees/main?recursive=1", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_osworld_g_jedi.py", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"], "focus": ["desktop GUI grounding", "infeasible-target refusal"], "notes": "The repository retains original OSWorld-G.json and refined OSWorld-G_refined.json instructions; treat them as alternative benchmark versions, not additional disjoint tasks. Refined instructions reduce requirements beyond grounding. The same repository also contains Jedi training/rendering material, which is excluded from benchmark counts. The standalone generic eval.py entry point is a stub; the named model runner supplies image_size to its geometry helper. No task count or successful execution is inferred from repository file totals. Apache-2.0 is recorded for code only.", "observations": ["screenshot", "instruction"], "setting": "offline"}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "e814851bc3016a2abf59358f99b80afd9c2bd5894fc34baddfb7e4d0aa03bd45", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "48308bba4f26b02cb7b8fbbc433bb31ec81c494dabd3af437ddbb10b8d9ce667", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/xlang-ai/OSWorld-G/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "64f64e0b847e0fa200aa6051e20b8035a436df00e2ef0d769688ab13fa4d3046", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_osworld_g_jedi.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "b85f5685ce6c200ea648471666177859693d7f40cf50c12d08765ae68580de4a", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/LICENSE"}], "id": "github:xlang-ai/osworld-g", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/OSWorld-G"}], "name": "OSWorld-G", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/README.md", "target_id": "arxiv:2505.13227", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Computer-use grounding benchmark with original and refined instructions, released with the Jedi synthetic data pipeline, a 4-million-example Jedi dataset and Jedi-3B and Jedi-7B models.", "topics": ["evaluation", "grounding"], "updated_at": null, "url": "https://github.com/xlang-ai/OSWorld-G"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["PyAutoGUI mouse and keyboard actions"], "counts": [{"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/evaluation_examples/test_all.json", "population": "evaluation_examples/test_all.json", "qualifier": "exact", "unit": "task definitions", "value": 494}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md", "population": "README task-suite coverage", "qualifier": "exact", "unit": "applications", "value": 20}], "evaluation": "Task-specific evaluation functions are invoked after agent execution.", "evidence_urls": ["https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md", "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/evaluation_examples/test_all.json"], "focus": ["data-science workflows", "multi-application task completion"], "notes": "Counted unique task IDs in the full manifest. Verbose versus abstract instructions and account versus non-account tasks form different subsets. The Ubuntu VM and external-account requirements are part of the protocol; no accounts or environments were provisioned.", "observations": ["screenshots", "accessibility trees", "set-of-marks views"], "setting": "interactive"}}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "e3edf30ab45da16382b46d8c542607c1bbbfcf077b6df57a2decf938fc84ff0f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "e3edf30ab45da16382b46d8c542607c1bbbfcf077b6df57a2decf938fc84ff0f", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:06:30Z", "role": "primary", "sha256": "338dd38ba7047b0cf6dbc68f3f7b9374f962708fa2dadcc8d126c232dfc31746", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/evaluation_examples/test_all.json"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "a8d1d88a0f9cce3cdbff73b03a80dc5e5d69e3ba2443946c268815a550c9e0ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/LICENSE"}], "id": "github:xlang-ai/spider2-v", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/Spider2-V"}], "name": "Spider2-V", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md", "target_id": "arxiv:2407.10956", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Visual agent evaluation on data-science and engineering workflows.", "topics": ["evaluation", "workflow"], "updated_at": null, "url": "https://github.com/xlang-ai/Spider2-V"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}], "id": "github:yiye3/guicourse", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/yiye3/GUICourse"}], "name": "GUICourse", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Three datasets for turning general VLMs into GUI agents: GUIEnv for OCR and grounding, GUIAct with 67k single-step and 15k multi-step web and Android actions, and GUIChat dialogues.", "topics": ["data-generation"], "updated_at": null, "url": "https://github.com/yiye3/GUICourse"}
{"aliases": [], "authors": [], "details": {}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "c984f114df33af953a97d22e14e14246f8a98e0866edcaf283900b309b71d8d1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/yjyddq/RiOSWorld/HEAD/README.md"}], "id": "github:yjyddq/riosworld", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/yjyddq/RiOSWorld"}], "name": "RiOSWorld", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/yjyddq/RiOSWorld/HEAD/README.md", "target_id": "arxiv:2506.00618", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Measures risky behavior of multimodal computer-use agents in an Ubuntu virtual machine, with risk scenarios such as phishing emails and released evaluation trajectories.", "topics": ["evaluation"], "updated_at": null, "url": "https://github.com/yjyddq/RiOSWorld"}
{"aliases": [], "authors": [], "details": {"platform_constraints": {"harmonyos": "The README limits device debugging support to HarmonyOS 4 or earlier."}}, "environments": ["android", "harmonyos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:14:29Z", "role": "primary", "sha256": "02e231c68ffc37c829789eabc508a220f3d99a33ee8bf676e53c999a0ff54021", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/Yuqi-Zhou/CHOP/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T13:29:25Z", "role": "primary", "sha256": "1a9f9af82764fdfc07dd8c2ec02e66cbb5c7c2cfc899b0e7b53b26e2a774ddf9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Yuqi-Zhou/CHOP/HEAD/LICENSE"}], "id": "github:yuqi-zhou/chop", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T13:14:29Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T13:29:25Z", "url": "https://raw.githubusercontent.com/Yuqi-Zhou/CHOP/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Yuqi-Zhou/CHOP"}], "name": "CHOP", "notes": "The README says the complete CHOP-ZH dataset remains unreleased. Android and HarmonyOS compatibility is publisher-documented and has not been runtime-tested here.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Yuqi-Zhou/CHOP/HEAD/README.md", "target_id": "arxiv:2503.03743", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile automation framework that uses reusable basis subtasks to constrain and refine planning between user goals and executable actions.", "topics": ["mobile-navigation", "planning", "workflow"], "updated_at": null, "url": "https://github.com/Yuqi-Zhou/CHOP"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "2d64349b514c43fb9d0260247300412d33cadc2a029d8279a00cb2a12f26a25a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:43Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/LICENSE"}], "id": "github:yuqi-zhou/gui-g1", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:43Z", "url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Yuqi-Zhou/GUI-G1"}], "name": "GUI-G1", "notes": "The README retains an older release-pending notice alongside a later checkpoint-release announcement. Check the linked model card for current weight availability and terms.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/README.md", "target_id": "arxiv:2505.15810", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI-grounding training code that modifies reasoning templates, reward functions, and the GRPO objective, with a linked model release.", "topics": ["grounding", "reinforcement-learning"], "updated_at": null, "url": "https://github.com/Yuqi-Zhou/GUI-G1"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending", "population": "AMEX collection described in the paper", "qualifier": "more-than", "reference_url": "https://arxiv.org/abs/2407.17490", "unit": "screenshots", "value": 104000}], "evidence_urls": ["https://raw.githubusercontent.com/YuxiangChai/AMEX-codebase/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"], "formats": [], "modalities": ["screenshots", "element annotations", "functionality descriptions", "action chains"], "notes": "The main repository contains companion training code. The scale and annotation levels here come from the paper, not an independent download or completeness audit.", "splits": [], "supervision": ["element grounding", "UI functionality", "interaction trajectories"]}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "3184bfd612543621e57226e9079132dc8d5695339795a60366d3f6d445e13551", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/YuxiangChai/AMEX-codebase/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["details.dataset_profile", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:11Z", "role": "primary", "sha256": "4aab61e6a95ab6b6598fb64b8e55d787cd49b5c4f242d15372aed2dd312f97aa", "supports": ["relations"], "url": "https://yxchai.com/AMEX/"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:02Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/YuxiangChai/AMEX-codebase/HEAD/LICENSE"}], "id": "github:yuxiangchai/amex-codebase", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:02Z", "url": "https://raw.githubusercontent.com/YuxiangChai/AMEX-codebase/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/YuxiangChai/AMEX-codebase"}], "name": "AMEX", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://yxchai.com/AMEX/", "target_id": "arxiv:2407.17490", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Code for AMEX Android GUI data that converts its three annotation levels into question-answer pairs and fine-tunes a LLaMA2-Accessory model, with a released checkpoint.", "topics": ["data-generation"], "updated_at": null, "url": "https://github.com/YuxiangChai/AMEX-codebase"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "de9c289c418218d40cdf1731775158af771d010266d7ea2aa398f2dc91f22acc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/zai-org/Open-AutoGLM/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "af69329b4667040c498cc87dfcc165db1f5f9452fe2ad3024ad7eef63d467666", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/zai-org/Open-AutoGLM/HEAD/LICENSE"}], "id": "github:zai-org/open-autoglm", "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/zai-org/Open-AutoGLM/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/zai-org/Open-AutoGLM"}], "name": "Open-AutoGLM", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Phone agent model and framework with device interaction tooling.", "topics": ["mobile-navigation"], "updated_at": null, "url": "https://github.com/zai-org/Open-AutoGLM"}
{"aliases": [], "authors": [], "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:17:23Z", "role": "primary", "sha256": "fc3009b355e46549c5128c933bb333aa37b1db5ac8ea8f2d0f300848d8a62200", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ZJU-ACES-ISE/UIDefectMocker/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T13:29:25Z", "role": "primary", "sha256": "2b75b184f28b204d8e2c7b45a341e4583202cf4c9da8964f108770929d1193ba", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ZJU-ACES-ISE/UIDefectMocker/HEAD/LICENSE"}], "id": "github:zju-aces-ise/uidefectmocker", "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T13:17:23Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T13:29:25Z", "url": "https://raw.githubusercontent.com/ZJU-ACES-ISE/UIDefectMocker/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZJU-ACES-ISE/UIDefectMocker"}], "name": "UI Defect Mocker", "notes": "This entry describes the released screenshot-defect tool, separately from the under-construction ChatUITest project hub.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Synthesizes visual UI defects from mobile screenshots and Android layout metadata to support GUI-testing experiments.", "topics": ["data-generation", "evaluation", "screen-understanding"], "updated_at": null, "url": "https://github.com/ZJU-ACES-ISE/UIDefectMocker"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "3f0aaefce32210dfba428fcf9e609e04e75d4d1520d24a3a6377611a9b83fa6d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ZJU-REAL/ClawGUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "af69329b4667040c498cc87dfcc165db1f5f9452fe2ad3024ad7eef63d467666", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ZJU-REAL/ClawGUI/HEAD/LICENSE"}], "id": "github:zju-real/clawgui", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/ZJU-REAL/ClawGUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZJU-REAL/ClawGUI"}], "name": "ClawGUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ZJU-REAL/ClawGUI/HEAD/README.md", "target_id": "arxiv:2604.11784", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Framework for GUI agent training, evaluation, and deployment.", "topics": ["reinforcement-learning", "evaluation"], "updated_at": null, "url": "https://github.com/ZJU-REAL/ClawGUI"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f9ecdf208ed4b7845ea4dbd1a7c936c016d11f7ec28f17e71338ab3c2f5b5c0d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-G2/HEAD/README.md"}], "id": "github:zju-real/gui-g2", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZJU-REAL/GUI-G2"}], "name": "GUI-G2", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-G2/HEAD/README.md", "target_id": "arxiv:2507.15846", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI grounding models at 3B and 7B trained with Gaussian rewards for click proximity to element centers and region coverage, with variance scaled to element size.", "topics": ["grounding"], "updated_at": null, "url": "https://github.com/ZJU-REAL/GUI-G2"}
{"aliases": [], "authors": [], "details": {"availability": "Public region-voting evaluation and policy-optimization source; model/data dependencies external.", "license_components_note": "Apache-2.0 notices in VLM-R1 subdirectories do not establish a root-project license.", "license_review": {"code": {"evidence_urls": ["https://api.github.com/repos/ZJU-REAL/GUI-RCPO/git/trees/main?recursive=1", "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/LICENSE", "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/src/open-r1-multimodal/LICENSE"], "reason": "Only subdirectory Apache-2.0 notices were established; no blanket project-source declaration is inferred."}}, "repository_revision": "af15ed5fe8b89b0fe5043a3e94f2984c7b126a4b"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "41c235305d28f6ba77ba1c5d1d9298fd6b1bd68ae3c9ea97eb34a8d22d0ba1dc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/zju-real/gui-rcpo/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "f5823de05f4828a8ab367ef13ec3a8338e020f0876e153ac2e944e4eb0e242c0", "supports": ["identity"], "url": "https://api.github.com/repos/zju-real/gui-rcpo"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "98ed8f0fa50699b8a3d0e8a71cd008113009a3072b971b8fadf2b92f68877e1e", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/ZJU-REAL/GUI-RCPO/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["notes", "licenses.code"], "url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["notes", "licenses.code"], "url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/src/open-r1-multimodal/LICENSE"}], "id": "github:zju-real/gui-rcpo", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZJU-REAL/GUI-RCPO"}], "name": "GUI-RC / GUI-RCPO", "notes": "The tree contains GUI-RC evaluation code and VLM-R1-based training scripts, with example JSONL-format paths. It requires separately supplied models and benchmark images; no released trained checkpoint is established here. Apache-2.0 files exist inside VLM-R1 subdirectories, but no repository-wide license is present for all newly added code, so the code slot remains unknown and the component notices are retained.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/zju-real/gui-rcpo/HEAD/README.md", "target_id": "arxiv:2508.05615", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI-grounding research code for spatial voting across sampled predictions and test-time policy optimization using region-consistency rewards.", "topics": ["grounding", "reinforcement-learning", "efficiency"], "updated_at": null, "url": "https://github.com/ZJU-REAL/GUI-RCPO"}
{"aliases": [], "authors": [], "details": {"availability": "Public WebArena/Mind2Web pipeline, workflow induction and inference source.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/LICENSE"], "reason": "Project source LICENSE, without transferring rights to external benchmark data or model services."}}, "repository_revision": "8c0ff8cd11d648c8fceb99e4e42f37e3b75381b1"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:56Z", "role": "primary", "sha256": "dd2cc4823b2a0b913f44fbaecf9f2cea98257871bdf974de6979459e49cb7b63", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:56Z", "role": "primary", "sha256": "f5c44d559932a22add3e94727b9b413d7c286f835db1396ae0fa1f05e20e773e", "supports": ["identity"], "url": "https://api.github.com/repos/zorazrw/agent-workflow-memory"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "602e2238829ad29ef15c75ac7dc61f8bd3dee6a1a74adf0d8ce023067fde6da3", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/zorazrw/agent-workflow-memory/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "1e504b3bd326f7858b39e7f2bedf5ff55d69ffb126e11af364b4184fb980882e", "supports": ["notes", "release_status"], "url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/webarena/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:11Z", "role": "primary", "sha256": "e047de93e1c73a39ab6bbf310f02e6023d15c2b06a2b1e80ad4a0875181a9533", "supports": ["notes", "release_status"], "url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/mind2web/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/LICENSE"}], "id": "github:zorazrw/agent-workflow-memory", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:56Z", "licenses": {"code": {"basis": "Project source LICENSE, without transferring rights to external benchmark data or model services.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/zorazrw/agent-workflow-memory"}], "name": "Agent Workflow Memory implementation", "notes": "The repository releases separate WebArena and Mind2Web pipelines for online or offline workflow induction and use. WebArena execution requires separately configured browser environments, while Mind2Web workflows require the original dataset and rankings. Publisher performance claims were not reproduced. Apache-2.0 is recorded for this implementation; external benchmark, model and service terms remain separate.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/README.md", "target_id": "arxiv:2409.07429", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Research code that extracts reusable browser-task routines from demonstrations or prior executions, stores them as workflow memory and reuses them during subsequent navigation.", "topics": ["workflow", "memory", "web-navigation"], "updated_at": null, "url": "https://github.com/zorazrw/agent-workflow-memory"}
{"aliases": [], "authors": [], "details": {"availability": "Public preprocessing/refinement code, offline action-matching evaluation, and multipart annotation archives; original images and a fine-tuned model are separate.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/LICENSE"], "reason": "Repository GNU GPL version 3 license for code; no data or pretrained-weight license inferred."}}, "repository_revision": "b02e8efdbe02ab82169a2c54d724e33b4037330e"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:55.267942+00:00", "role": "primary", "sha256": "77384ff1dfceac74f2e60af56b016a815963481f9e41a29915698777f55e828e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T11:25:27+00:00", "role": "primary", "sha256": "3d57dc43db78620832e41416e4bfb7ae445afe782e5538d85a9784a8f1e5e4eb", "supports": ["identity"], "url": "https://api.github.com/repos/ZrW00/GUIPivot"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:54.310430+00:00", "role": "primary", "sha256": "a4d3ddbb2b2f4a936c9bc1a1c388b9870dbc5d2c61ff6841a1082b884f5c54e9", "supports": ["release_status", "details.repository_revision", "details.availability", "notes"], "url": "https://api.github.com/repos/ZrW00/GUIPivot/git/trees/b02e8efdbe02ab82169a2c54d724e33b4037330e?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:03:28.611747+00:00", "role": "primary", "sha256": "3972dc9744f6499f0f9b2dbf76696f2ae7ad8af9b23dde66d6af86c9dfb36986", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:36.501726+00:00", "role": "primary", "sha256": "25e5282e7d85bca133fc77e294cd95d90e4ae65b0d8796cd7c49f53a633b4893", "supports": ["summary", "environments", "notes"], "url": "https://arxiv.org/html/2503.00401v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:03:30.911159+00:00", "role": "primary", "sha256": "9b5b196e065be2ebc05169a68710b8a6e1bd3b6eb663c68bb449234dbb7a45af", "supports": ["notes", "details.availability"], "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/data/dataset_info.json"}], "id": "github:zrw00/guipivot", "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T12:02:55.267942+00:00", "licenses": {"code": {"basis": "Repository GNU GPL version 3 license for code; no data or pretrained-weight license inferred.", "declared": "GPL-3.0", "retrieved_at": "2026-10-18T12:03:28.611747+00:00", "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZrW00/GUIPivot"}], "name": "GUIPivot", "notes": "The repository releases preprocessing/refinement/evaluation code and multipart ShareGPT-format JSON annotations. Evaluation matches recorded single-step action predictions on AndroidControl and AITZ; it does not run live device tasks. The README links a UIBERT training-JSON path that is not a standalone file in this revision; annotations are packaged in data/GUI.z01, data/GUI.z02 and data/GUI.zip. Original UIBERT/OS-Atlas, AITZ and AndroidControl images must be obtained separately. This entry does not claim a released trained checkpoint or transfer the GPL-3.0 code license to third-party datasets and images.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/README.md", "target_id": "arxiv:2503.00401", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Query-inference training tools connecting screenshot grounding with GUI action reasoning, including preprocessing, query refinement and offline single-step evaluation on AndroidControl and AITZ.", "topics": ["grounding", "planning", "imitation-learning"], "updated_at": null, "url": "https://github.com/ZrW00/GUIPivot"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "bd3e99c1e12588bc3136dfa27de2c0d5790b6f6f25f2696988eb182c6b1ef670", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://claude.com/chrome"}], "id": "organization:anthropic", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://claude.com/chrome"}], "name": "Anthropic", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with Anthropic computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://claude.com/chrome"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "df8615d0687cba27e5558ea3f7ecb1823871e3804522603dd62b52dcf127ac89", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.askui.com/"}], "id": "organization:askui", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.askui.com/"}], "name": "AskUI", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [{"evidence_url": "https://www.askui.com/", "target_id": "organization:askui", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with AskUI computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://www.askui.com/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "69e029d01dda3806f88abfaee01da08c671adcb3555f1314e975ab8942a686ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://browser-use.com/"}], "id": "organization:browser-use", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://browser-use.com/"}], "name": "Browser Use", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [{"evidence_url": "https://browser-use.com/", "target_id": "organization:browser-use", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with Browser Use computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://browser-use.com/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "463db02b607e7215a4421440c4d274ee3546bd9eefc50ed9e10f64d27d7f8f38", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.browseract.com/"}], "id": "organization:browseract", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browseract.com/"}], "name": "BrowserAct", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [{"evidence_url": "https://www.browseract.com/", "target_id": "organization:browseract", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with BrowserAct computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://www.browseract.com/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "d86315819750e4a5fde12bce85a03c61cf68e704cb54146e6180027a37471b72", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.browserbase.com/"}], "id": "organization:browserbase", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browserbase.com/"}], "name": "Browserbase", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with Browserbase computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://www.browserbase.com/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "e044e75ec1c2a74f2bee8c02b6efae05c399ec3537e34dd861dcf4a0a25dc182", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.browserless.io/"}], "id": "organization:browserless", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browserless.io/"}], "name": "Browserless", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [{"evidence_url": "https://www.browserless.io/", "target_id": "organization:browserless", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with Browserless computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://www.browserless.io/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "6232364481bbae45a3fcec1d1aca858493baff0693ececa98b09b610183e6796", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://e2b.dev/"}], "id": "organization:e2b", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://e2b.dev/"}], "name": "E2B", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [{"evidence_url": "https://e2b.dev/", "target_id": "organization:e2b", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with E2B computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://e2b.dev/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "e7062cfb4c0948d3a2274843f9f493f34cb40dfd634371408086df6141a0e036", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://ai.google.dev/gemini-api/docs/computer-use"}], "id": "organization:google", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://ai.google.dev/gemini-api/docs/computer-use"}], "name": "Google", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with Google computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://ai.google.dev/gemini-api/docs/computer-use"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "84ce0fd682367a6dac5cc5abc39d5c877556024bbf6ffdad30f76fc2ec08523f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://hcompany.ai/research/models"}], "id": "organization:h", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://hcompany.ai/research/models"}], "name": "H", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with H computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://hcompany.ai/research/models"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "47f31d41c806e97143fa046c4e1f5e916096d158bee0bff43efd15a7d78053b1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.hyperbrowser.ai/"}], "id": "organization:hyperbrowser", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.hyperbrowser.ai/"}], "name": "Hyperbrowser", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [{"evidence_url": "https://www.hyperbrowser.ai/", "target_id": "organization:hyperbrowser", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with Hyperbrowser computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://www.hyperbrowser.ai/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "832b84c366eaadc69e328f207f2659cf07a0901c47662244831c0f1695d02efd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://manus.im/"}], "id": "organization:manus", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://manus.im/"}], "name": "Manus", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [{"evidence_url": "https://manus.im/", "target_id": "organization:manus", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with Manus computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://manus.im/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "14f573af0c4ea3e5922b8831b7f03724936963e293d7c0aaa3fccbb71dbb0cfa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://openai.com/index/computer-using-agent/"}], "id": "organization:openai", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/computer-using-agent/"}], "name": "OpenAI", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with OpenAI computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://openai.com/index/computer-using-agent/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "4cb358ab96e57a66d2b268a17963f92be249325fb0687a78985a6d9773ff5a81", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.simular.ai/"}], "id": "organization:simular", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.simular.ai/"}], "name": "Simular", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [{"evidence_url": "https://www.simular.ai/", "target_id": "organization:simular", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with Simular computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://www.simular.ai/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "7a8813c174043d1a27cb3766f77c90beb1a08eeedee935264d6237d6a57672a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.skyvern.com/"}], "id": "organization:skyvern", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.skyvern.com/"}], "name": "Skyvern", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [{"evidence_url": "https://www.skyvern.com/", "target_id": "organization:skyvern", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with Skyvern computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://www.skyvern.com/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "5eacfa15c116ddeb02f6fd731fbb2e9856a933d743870fe41d947ef3b01aeb6f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://steel.dev/"}], "id": "organization:steel", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://steel.dev/"}], "name": "Steel", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with Steel computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://steel.dev/"}
{"aliases": [], "authors": [], "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "461de0060fae1ded024dfa724624b001665b463b14f454afafebd0c9a2dce71d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.tinyfish.ai/"}], "id": "organization:tinyfish", "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.tinyfish.ai/"}], "name": "TinyFish", "notes": "Publisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.", "published_at": null, "relations": [{"evidence_url": "https://www.tinyfish.ai/", "target_id": "organization:tinyfish", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Publisher or product brand associated with TinyFish computer-use or browser automation offerings.", "topics": ["industry"], "updated_at": null, "url": "https://www.tinyfish.ai/"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/tree/main", "gated": false, "hub_id": "Tongyi-MAI/MAI-UI-8B"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "a1087d2a0dde6e7ba6d79db546ee6360a5ffd95ffbc3add1478eedec603a4187", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "349e6d5753fa94a47feb4b2fc53ecd6b9188b645a0cffabef92fd1aa0976ecad", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Tongyi-MAI/MAI-UI-8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "a1087d2a0dde6e7ba6d79db546ee6360a5ffd95ffbc3add1478eedec603a4187", "supports": ["licenses.weights"], "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md"}], "id": "url:00b1a3abf9bcb424b264", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B"}], "name": "MAI-UI 8B", "notes": "Four public safetensors shards are present and Apache-2.0 is declared. The card presents family-wide results and device-cloud architecture; those claims are not all assigned to the 8B checkpoint. Other parameter sizes are not duplicated in this proposal batch.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md", "target_id": "arxiv:2512.22047", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A released MAI-UI checkpoint combining mobile GUI actions, agent-user interaction and tool-use-oriented navigation.", "topics": ["grounding", "mobile-navigation", "tool-integration"], "updated_at": null, "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/henryhe0123/PC-Agent-E/tree/568ab45861d5dab041e158d47211c6f53ba343f2", "gated": false, "hub_id": "henryhe0123/PC-Agent-E", "hub_revision": "568ab45861d5dab041e158d47211c6f53ba343f2", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/henryhe0123/PC-Agent-E/raw/568ab45861d5dab041e158d47211c6f53ba343f2/README.md"], "reason": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate."}}, "repository_url": "https://github.com/GAIR-NLP/PC-Agent-E"}, "environments": ["desktop", "windows", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "feb41002efff5ef30fe06f9891efbe954e0d046c0379f7c8ca72b5d8a8a304be", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.weights"], "url": "https://huggingface.co/henryhe0123/PC-Agent-E/raw/568ab45861d5dab041e158d47211c6f53ba343f2/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:21Z", "role": "primary", "sha256": "f81a6cea4234e65f8bab390a18487e65dd0b0cb62f0a9458c658ff07e8a7b5f7", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url"], "url": "https://huggingface.co/api/models/henryhe0123/PC-Agent-E"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:13Z", "role": "primary", "sha256": "3ffb8c843c6acb8c2d5bbba11ba49d6b36cd566553a89974b6252a2bfd289d5d", "supports": ["environments", "details.repository_url"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md"}], "id": "url:04aa7569d183379d615c", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-18T04:35:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://huggingface.co/henryhe0123/PC-Agent-E/raw/568ab45861d5dab041e158d47211c6f53ba343f2/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/henryhe0123/PC-Agent-E"}], "name": "PC-Agent-E checkpoint", "notes": "Public manifest lists 31 Safetensors shards and their index. The card prose names Qwen2.5-VL-72B-Instruct while its YAML base_model field self-references PC-Agent-E; no base-model relationship is imported from that inconsistent metadata. The model-card MIT declaration is recorded for weights only.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/henryhe0123/PC-Agent-E/raw/568ab45861d5dab041e158d47211c6f53ba343f2/README.md", "target_id": "arxiv:2505.13909", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Computer-use checkpoint fine-tuned on PC-Agent-E trajectories; the model-card prose identifies Qwen2.5-VL-72B-Instruct as its starting model.", "topics": ["action-prediction", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/henryhe0123/PC-Agent-E"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "47f31d41c806e97143fa046c4e1f5e916096d158bee0bff43efd15a7d78053b1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.hyperbrowser.ai/"}], "id": "url:053ebfe0afac66c86222", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.hyperbrowser.ai/"}], "name": "Hyperbrowser", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://www.hyperbrowser.ai/", "target_id": "organization:hyperbrowser", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Browser infrastructure and automation services for AI agents.", "topics": ["sandbox"], "updated_at": null, "url": "https://www.hyperbrowser.ai/"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md", "https://huggingface.co/api/datasets/xlangai/aguvis-stage1", "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md", "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage1.yaml"], "formats": ["JSON annotations", "ZIP image packages", "multipart TAR image packages"], "modalities": ["GUI screenshots", "grounding annotations"], "notes": "Stage 1 grounding release, Hub revision 5d11f8f5. JSON annotations and ZIP/multipart TAR image packages are publicly listed. The official training configuration groups web, desktop and mobile inputs and selects all rows of its configured inputs; this is a training recipe, not a release sample count or held-out split. The config names omniact.json while the Hub manifest lists omniact_fix.json, so local paths need reconciliation. No sample total is inferred from names such as webui350k. Apache-2.0 is the release card declaration; upstream source rights remain separate.", "splits": ["training"], "supervision": ["GUI grounding supervised fine-tuning"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/tree/main", "gated": false, "hub_id": "xlangai/aguvis-stage1"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "38bda302cf741b9d2b2394b53d157b3c19d8a73e3704c01ed2f9a27c374feb77", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "details.dataset_profile"], "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "cb3b8ec88d811d946df3cbf38ef555db5a10d1124668765710c380e9b1988b06", "supports": ["identity", "release_status", "details", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/xlangai/aguvis-stage1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dd9f36b818c1b7b6bd772ba8575f6c0ebab6d9e0d307edcc2f196f3e4ba502e", "supports": ["environments", "summary", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:11Z", "role": "primary", "sha256": "f6a9e55c5d48469e2c8f7799d917d25e7a758036f71c7c045fa92c35057d20da", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage1.yaml"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "38bda302cf741b9d2b2394b53d157b3c19d8a73e3704c01ed2f9a27c374feb77", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md"}], "id": "url:06dd5ac7917c45d48509", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1"}], "name": "Aguvis stage 1 grounding data", "notes": "The manifest contains grounding JSON annotations and packaged images. The card declares Apache-2.0, which is recorded for this release without overriding source-dataset rights. This first-stage grounding collection is distinct from the project’s second-stage trajectory collection. No archive or training sample was downloaded.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md", "target_id": "arxiv:2412.04454", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The first-stage Aguvis training collection, packaging GUI element-grounding annotations and image archives from multiple interface-data sources.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "population": "context-aware episode training release", "qualifier": "approximately", "unit": "instruction-output pairs", "value": 992000}], "evidence_urls": ["https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"], "formats": [], "modalities": ["screen observations", "instructions", "interaction history"], "notes": "The Hub card contains only licensing metadata. The project README supports the episode-training description and rounded count; archive-level quality and exact row counts remain unverified.", "splits": [], "supervision": ["context-aware grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data/tree/main", "gated": false, "hub_id": "Aria-UI/Aria-UI_Context-aware_Data", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "79f91aab71bb8a0ec3ad9859cbc42d70227e01db729ec80dc5690b466b259cad", "supports": ["name", "kind"], "url": "https://huggingface.co/api/datasets/Aria-UI/Aria-UI_Context-aware_Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a2a3c5ad115e4e2109b56686c9e0b073bcfec6d9f052a58bc92683af44933480", "supports": ["name", "summary", "environments", "topics", "scope", "details.dataset_profile", "relations"], "url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "79f91aab71bb8a0ec3ad9859cbc42d70227e01db729ec80dc5690b466b259cad", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Aria-UI/Aria-UI_Context-aware_Data"}], "id": "url:07b65dc937756ddff4a1", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:11:53Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T11:11:53Z", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data"}], "name": "Aria-UI Context-aware Data", "notes": "The publisher reports approximately 992K instruction-output pairs in its repository. The Hub card contains only a license declaration; schema quality and exact sample counts were not independently verified.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "target_id": "github:ariaui/aria-ui", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "Episode-based mobile GUI grounding data released with Aria-UI for incorporating interaction history into target localization.", "topics": ["grounding", "memory"], "updated_at": null, "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "bd3e99c1e12588bc3136dfa27de2c0d5790b6f6f25f2696988eb182c6b1ef670", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://claude.com/chrome"}], "id": "url:07b6bbce7ce26a847f79", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://claude.com/chrome"}], "name": "Claude for Chrome", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://claude.com/chrome", "target_id": "organization:anthropic", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Claude browser extension for interacting with websites.", "topics": ["web-navigation"], "updated_at": null, "url": "https://claude.com/chrome"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Aria-UI/Aria-UI-context-aware/tree/main", "gated": false, "hub_id": "Aria-UI/Aria-UI-context-aware", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["android", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "183a8de8928b3b71b171a35149621de078aaf80b4ebde377a5dc3b5cd4a04a27", "supports": ["name", "kind", "summary", "environments", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/Aria-UI/Aria-UI-context-aware"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a2a3c5ad115e4e2109b56686c9e0b073bcfec6d9f052a58bc92683af44933480", "supports": ["summary", "environments", "relations", "release_status", "topics"], "url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "183a8de8928b3b71b171a35149621de078aaf80b4ebde377a5dc3b5cd4a04a27", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/Aria-UI/Aria-UI-context-aware"}], "id": "url:085b6811ddd6f797475d", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/Aria-UI/Aria-UI-context-aware"}], "name": "Aria-UI context-aware checkpoint", "notes": "The Hub manifest exposes twelve safetensors shards and model code/configuration, but no README card or model-license declaration. The official repository explicitly announces this context-aware release separately from the base model. Code or base-model licenses are not inherited.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "target_id": "arxiv:2412.16256", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The context-aware Aria-UI release that uses interaction context to support grounding in dynamic Android and desktop agent tasks.", "topics": ["grounding", "interaction-history", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/Aria-UI/Aria-UI-context-aware"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "GUIEnv-local SFT data", "qualifier": "approximately", "unit": "region-text QA pairs", "value": 700000}], "evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"], "formats": ["JSON", "Parquet"], "modalities": ["screenshots", "text", "bounding boxes"], "notes": "Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.", "splits": ["train", "test"], "supervision": ["OCR", "region-text grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/yiye2023/GUIEnv/tree/main", "gated": false, "hub_id": "yiye2023/GUIEnv", "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "https://huggingface.co/datasets/yiye2023/GUIEnv/raw/main/README.md"], "reason": "Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "15813be05dc76aed5f667875188bf9612e74ff2f4699d270dd0d1aeda86767f8", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/yiye2023/GUIEnv"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "4bcf87ecfbbb8e07a01b21415a970c8b53a5283bf6872b657040d3f45c9241f7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yiye2023/GUIEnv/raw/main/README.md"}], "id": "url:0af07e00658a430cee6a", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yiye2023/GUIEnv"}], "name": "GUIEnv-local", "notes": "Public JSON annotations and Parquet image packages are present. The project README declares CC-BY-4.0 for datasets, while this Hub card declares Apache-2.0; these conflicting declarations are retained without asserting a single confirmed data license. The project code’s MIT declaration is separate. Only GUIEnv-local is marked released; the advertised 10M-example GUIEnv-global pretraining collection is still unchecked and is not counted here.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The released local-supervision portion of GUIEnv, pairing webpage regions and text for OCR and GUI grounding.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/yiye2023/GUIEnv"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://www.interactionmining.org/archive/rico", "population": "original Rico downloads described on the project page", "qualifier": "more-than", "unit": "unique UI screens", "value": 66000}, {"basis": "reported", "evidence_url": "https://www.interactionmining.org/archive/rico", "population": "original Rico collection described on the project page", "qualifier": "more-than", "unit": "Android apps", "value": 9300}, {"basis": "reported", "evidence_url": "https://www.interactionmining.org/archive/rico", "population": "crowd-worker collection before automated exploration", "qualifier": "exact", "unit": "human interaction traces", "value": 10811}], "evidence_urls": ["https://www.interactionmining.org/archive/rico", "https://danafergan.com/publications/deka2017rico.pdf"], "formats": ["PNG", "JSON", "CSV", "NumPy NPY", "GIF"], "modalities": ["screenshots", "Android view hierarchies", "gesture traces", "transition animations", "app/UI metadata"], "notes": "Project-page scope: screenshots paired with augmented view hierarchies, gesture coordinates, layout vectors and metadata. These observations are not action/task totals. The later semantic-annotation download is a related augmentation of screens, not an additional disjoint corpus. The retained original paper reports more than 72k screens from more than 9.7k apps while this project page says more than 66k screens and more than 9.3k apps; the difference is unresolved. Profile counts follow the project page and do not merge these populations. No train/validation/test partition or data license is established here; archives were not downloaded.", "splits": [], "supervision": ["UI structure and element properties", "recorded human interactions", "learned UI layout embeddings"]}, "modalities": ["screenshot", "view-hierarchy", "interaction-trace"], "reported_scale": {"comparison": ">", "unit": "unique UI screens", "value": 66000}, "sample_count": null, "sample_unit": "UI screens"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:29Z", "role": "primary", "sha256": "daa370b9aae5bce9f5171ea81fa4574c6155e04177af1db138529efd76377f89", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "notes", "details.dataset_profile"], "url": "https://www.interactionmining.org/archive/rico"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T10:01:32Z", "role": "primary", "sha256": "a1ec58c526b152c052c5c72f9a62725f1026b5f0e4f45c8fe85fce284295da84", "supports": ["notes", "details.dataset_profile"], "url": "https://danafergan.com/publications/deka2017rico.pdf"}], "id": "url:0b134b082a7bd507fc6a", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.interactionmining.org/archive/rico"}], "name": "Rico", "notes": "The official project page cites the 2017 Rico paper. The project page reports more than 66,000 screens, while the original publication reports more than 72,000. These source-specific counts are kept separate; the reason for the difference is not established by these sources. App counts also differ: the official project page reports more than 9.3k Android apps and the retained published abstract more than 9.7k. Profile counts use the project-page scope; no reconciliation is inferred.", "published_at": null, "relations": [{"evidence_url": "https://www.interactionmining.org/archive/rico", "target_id": "url:b560b5d6abaab98483f6", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile app screenshots, view hierarchies, and interaction traces for UI research.", "topics": ["screen-understanding"], "updated_at": null, "url": "https://www.interactionmining.org/archive/rico"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/vyokky/GUI-360/tree/main", "gated": false, "hub_id": "vyokky/GUI-360", "languages": [], "modalities": [], "size_category_reported": ["1M<n<10M"]}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "e51af2845c2f31c7af112579b42c3ff9bf5a2c2356e80932e829bc2be1a341bd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/vyokky/GUI-360"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "e51af2845c2f31c7af112579b42c3ff9bf5a2c2356e80932e829bc2be1a341bd", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/vyokky/GUI-360"}], "id": "url:0b590d03a3d5103a79e4", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/vyokky/GUI-360"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/vyokky/GUI-360"}], "name": "GUI-360", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/vyokky/GUI-360", "target_id": "arxiv:2511.04307", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Desktop GUI trajectory data supporting multiple interface understanding tasks.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/vyokky/GUI-360"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "4cb358ab96e57a66d2b268a17963f92be249325fb0687a78985a6d9773ff5a81", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.simular.ai/"}], "id": "url:0bdd6b5ebf2e1f248828", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.simular.ai/"}], "name": "Simular", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://www.simular.ai/", "target_id": "organization:simular", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Computer-use products and research from Simular.", "topics": ["workflow"], "updated_at": null, "url": "https://www.simular.ai/"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/tree/main", "gated": false, "hub_id": "SenseLLM/GUI-Lasagne-L1"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "270b3af0904a2a50d040948100a327458839dffa5083fc74bb030298f5ab3985", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "224d6487de47cc63251c3b5c5c9f6e75e692b5ef0e12f0a7f71000da0b866e6f", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/SenseLLM/GUI-Lasagne-L1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "916bde91ff00aba22ff13ed3d8b4ca679bc8c198632f1d3bc9469660287cb5d3", "supports": ["environments", "summary"], "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "270b3af0904a2a50d040948100a327458839dffa5083fc74bb030298f5ab3985", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md"}], "id": "url:0c5659ca9e59d942aea3", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1"}], "name": "GUI-Lasagne L1", "notes": "Public archive shards are present. The dataset card declares Apache-2.0; it does not resolve rights in third-party interface content. The short card supplies no reliable sample count or explicit train/test partition, so these remain unspecified. The dataset card and archive inventory do not establish its platform distribution; the associated model's platform coverage is not inherited. An associated SpiritSight model card still describes dataset release as forthcoming; the independent dataset card and public archive manifest establish this release's availability.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md", "target_id": "arxiv:2503.03196", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The published GUI-Lasagne training-data collection used for SpiritSight, distributed as dated archive shards for visual interface learning.", "topics": ["grounding", "screen-understanding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "population": "train split", "qualifier": "exact", "unit": "rows", "value": 63530}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "population": "validation split", "qualifier": "exact", "unit": "rows", "value": 7944}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "population": "test split", "qualifier": "exact", "unit": "rows", "value": 7938}], "evidence_urls": ["https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"], "formats": ["Parquet"], "modalities": ["images", "text", "bounding boxes", "OCR metadata"], "notes": "Card-reported row counts are not unique-screen counts. The source field is important for the per-image license terms; the compilation does not relicense original images.", "splits": ["train", "validation", "test"], "supervision": ["instruction-to-target boxes"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/agentsea/wave-ui/tree/main", "gated": false, "hub_id": "agentsea/wave-ui", "languages": [], "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"], "reason": "The compilation expressly does not relicense its images and lists differing source terms. No blanket permissive data license is appropriate."}}, "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "76e5e458eb8788f2c2cfddd06cb944327f363ec70120e122ce93d1350cf82e19", "supports": ["name", "kind", "identity", "release_status", "details", "summary"], "url": "https://huggingface.co/api/datasets/agentsea/wave-ui"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "27ec13aa77dc55113e4d1a19c2269070e2d69d2fe8df8e2313721a1e3fc66bd9", "supports": ["details"], "url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:57Z", "role": "primary", "sha256": "63c6b3ce6a3d9c21a505865ceed698df2894bcb292590c59686e9477df867942", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "e42bbcf8d466c977f9b2150feb73d8cbda2ff907c279630ad4730a8b1f5d665c", "supports": ["summary", "topics", "scope"], "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "76e5e458eb8788f2c2cfddd06cb944327f363ec70120e122ce93d1350cf82e19", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/agentsea/wave-ui"}], "id": "url:0dbb55f127da04d95222", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": {"basis": "Publisher compilation license; individual images keep their source-dataset terms and receive no additional rights from this compilation.", "declared": "Wave UI Dataset License; per-image source terms retained", "retrieved_at": "2026-10-17T04:06:57Z", "url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/agentsea/wave-ui"}], "name": "Wave UI", "notes": "Train, validation and test Parquet files are publicly listed. The compilation license explicitly preserves each image’s original source terms and grants no extra rights to those images. The card exposes a platform column without listing its values, so operating-system coverage remains unspecified.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Screenshot-instruction records with target boxes and UI metadata such as OCR, source, element type and platform.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/agentsea/wave-ui"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md", "population": "SeeClick-derived web subset", "qualifier": "more-than", "unit": "screenshots", "value": 270000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md", "population": "SeeClick-derived web subset", "qualifier": "more-than", "unit": "UI elements", "value": 3000000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md", "population": "FineWeb-derived subset", "qualifier": "at-least", "unit": "images", "value": 1600000}], "evidence_urls": ["https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md"], "formats": ["JSON annotations", "image archives"], "modalities": ["screenshots", "referring expressions", "bounding boxes"], "notes": "These are subset-specific counts, not a total for the full OS-Atlas release. Raw boxes use 0–1 coordinates; the documented training wrapper scales them to 0–1000.", "splits": [], "supervision": ["element grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Atlas-data", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "c2c5f7b3f07d92349ffd8256792599c856ed77c89108c45dbaef48a4699abe20", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Atlas-data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:40:46Z", "role": "primary", "sha256": "19f28b4bc0798e28180b968010c0f7c8dfed2dbfb24b122b958476b871d2e4fd", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "c2c5f7b3f07d92349ffd8256792599c856ed77c89108c45dbaef48a4699abe20", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Atlas-data"}], "id": "url:0fdeb1cf5f49f84820e6", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data"}], "name": "OS-Atlas Data", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Atlas-data", "target_id": "arxiv:2410.23218", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI grounding pre-training data for OS-Atlas in desktop (Linux, macOS, Windows), mobile and web domains, with element boxes stored as relative left, top, right and bottom ratios.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/biglab/webui-all/tree/main", "gated": false, "hub_id": "biglab/webui-all", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "70928db908760c1561bfc8e0cd7d925e0103ffddef4676fbaa7c1756a804b99a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/biglab/webui-all"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:53Z", "role": "primary", "sha256": "93bd04914735f748d687b71597970a79d95ae8636e7ed75d267ebd6954d11a8f", "supports": ["notes"], "url": "https://huggingface.co/datasets/biglab/webui-all/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "70928db908760c1561bfc8e0cd7d925e0103ffddef4676fbaa7c1756a804b99a", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/biglab/webui-all"}], "id": "url:13b9442bcb6a8b984c4f", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "other", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/biglab/webui-all"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/biglab/webui-all"}], "name": "WebUI", "notes": "The card links the WebUI project's ACM publication (DOI 10.1145/3544548.3581158) and website, not an arXiv record or a catalogued repository, so no paper or family relation is recorded.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Web screenshot and interface data for UI research.", "topics": ["screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/biglab/webui-all"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md", "https://huggingface.co/api/datasets/henryhe0123/PC-Agent-E"], "formats": ["ZIP package"], "modalities": ["GUI screenshots", "action trajectories", "generated reasoning"], "notes": "Public manifest exposes data.zip; the internal file schema and number of examples remain unverified. Supervision describes the publisher’s refinement/thought-completion/trajectory-boost pipeline, not an independently sampled data-quality audit.", "splits": ["training"], "supervision": ["human demonstration refinement", "generated action reasoning", "trajectory augmentation"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E/tree/4f64c7b055154bed9ffea677aa86aeb51a01cb73", "gated": false, "hub_id": "henryhe0123/PC-Agent-E", "hub_revision": "4f64c7b055154bed9ffea677aa86aeb51a01cb73", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/henryhe0123/PC-Agent-E/raw/4f64c7b055154bed9ffea677aa86aeb51a01cb73/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}, "repository_url": "https://github.com/GAIR-NLP/PC-Agent-E"}, "environments": ["desktop", "windows", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:13Z", "role": "primary", "sha256": "3ffb8c843c6acb8c2d5bbba11ba49d6b36cd566553a89974b6252a2bfd289d5d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:21Z", "role": "primary", "sha256": "66f70ee2704cdf2e8847cfe04be05d9e14b3a3c1bee953e744f8209d6c9c2b85", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/henryhe0123/PC-Agent-E"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:17Z", "role": "primary", "sha256": "4dfd1f3ec40f7ff3c94896955216cb60a25114b320caff29e909d296d375d6ef", "supports": ["relations", "licenses.data"], "url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E/raw/4f64c7b055154bed9ffea677aa86aeb51a01cb73/README.md"}], "id": "url:14449fe321e0a1ca34ea", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:27:13Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:17Z", "url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E/raw/4f64c7b055154bed9ffea677aa86aeb51a01cb73/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E"}], "name": "PC-Agent-E training data", "notes": "Hub release contains data.zip. The archive was not downloaded and its internal schema/example total were not independently inspected. Training-purpose and platform descriptions come from the official pipeline README; the brief card establishes the paper association and MIT data declaration.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md", "target_id": "arxiv:2505.13909", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Released dataset used to train PC-Agent-E, linked by the official pipeline for refining human trajectories and synthesizing action reasoning.", "topics": ["data-generation", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "7a8813c174043d1a27cb3766f77c90beb1a08eeedee935264d6237d6a57672a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.skyvern.com/"}], "id": "url:15ed5fd3cd950d837f98", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.skyvern.com/"}], "name": "Skyvern Cloud", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://www.skyvern.com/", "target_id": "organization:skyvern", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Hosted Skyvern service for AI browser workflows such as invoice downloads and form filling, with built-in CAPTCHA and 2FA handling, proxy geotargeting, run explanations and JSON or CSV output.", "topics": ["workflow"], "updated_at": null, "url": "https://www.skyvern.com/"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/tree/main", "gated": false, "hub_id": "Hcompany/Holo4-35B-A3B"}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "4f1aa5006aa9adaeb5ac549e5a5f8ea164352714a00528b1afb8602812d45b44", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "7a41bdbfb4a300302a7d3633f12962505d56ba874a37292afccbd48b1bec8287", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo4-35B-A3B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "4f1aa5006aa9adaeb5ac549e5a5f8ea164352714a00528b1afb8602812d45b44", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/raw/main/README.md"}], "id": "url:16083a0225a60882c704", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:11:45Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:11:45Z", "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B"}], "name": "Holo4 35B-A3B BF16", "notes": "The public manifest contains 26 safetensors shards for the BF16 checkpoint and the card declares Apache-2.0. Quantized FP8/NVFP4/GGUF variants are not split into duplicate entries here. The card also contains demos and scores for Holo4-27B; those are not assigned to this 35B-A3B checkpoint. No training corpus or runtime was downloaded.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A Qwen3.6-based computer-use checkpoint designed to request GUI actions, code execution and tool calls through an external agent harness.", "topics": ["grounding", "action-prediction", "tool-integration"], "updated_at": null, "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B"}
{"aliases": ["url:1ab38d1d6a4c46a8eb3e", "OpenMobile-2/OpenMobile-27B", "https://huggingface.co/OpenMobile-2/OpenMobile-27B"], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.6-27B", "gated": false, "hub_revision": "3d844db8c20421524a5012837b151e444ad35ae6", "listed_weight_files": ["model-00001-of-00012.safetensors", "model-00002-of-00012.safetensors", "model-00003-of-00012.safetensors", "model-00004-of-00012.safetensors", "model-00005-of-00012.safetensors", "model-00006-of-00012.safetensors", "model-00007-of-00012.safetensors", "model-00008-of-00012.safetensors", "model-00009-of-00012.safetensors", "model-00010-of-00012.safetensors", "model-00011-of-00012.safetensors", "model-00012-of-00012.safetensors"]}, "environments": ["android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "ba760ab75e997941dabf0845b34e5fb83daeb83a5cfa16b13929477cfc61a31c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "c7e388efe1ecfd415b7b5af7f296e62646a0a1380037f1e57ad1a43c26463b9a", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/OpenMobile-2/OpenMobile-2-27B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:06Z", "role": "primary", "sha256": "9b385e9f6aa2f7169f616a76887df3b76df605cb2bea071d99ee7ca3f9b659cc", "supports": ["environments"], "url": "https://os-copilot.github.io/OpenMobile2-Home/"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "ba760ab75e997941dabf0845b34e5fb83daeb83a5cfa16b13929477cfc61a31c", "supports": ["licenses.weights"], "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:11Z", "role": "primary", "sha256": "f669181b7f1b92bbbcb8ab71f0d15d8c3cf93b3665d73d20e9b9f0dc3a36e158", "supports": ["aliases"], "url": "https://huggingface.co/api/models/OpenMobile-2/OpenMobile-27B"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:12Z", "role": "primary", "sha256": "f669181b7f1b92bbbcb8ab71f0d15d8c3cf93b3665d73d20e9b9f0dc3a36e158", "supports": ["aliases"], "url": "https://huggingface.co/api/models/OpenMobile-2/OpenMobile-2-27B"}], "id": "url:1611de0e6d606165d12d", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:51Z", "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B"}], "name": "OpenMobile-2 27B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md", "target_id": "github:openmobile2/openmobile2-code", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md", "target_id": "url:546e6c95825d1318c98d", "type": "trained_on"}, {"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md", "target_id": "arxiv:2610.14157", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile-agent checkpoint based on Qwen3.6-27B for GUI and app-native tool interaction.", "topics": ["mobile-navigation", "tool-integration"], "updated_at": null, "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/UI-S1/README.md", "population": "public example subset", "qualifier": "exact", "unit": "training examples", "value": 1000}], "evidence_urls": ["https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md", "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/UI-S1/README.md"], "formats": ["JSONL", "Parquet"], "modalities": ["images", "text", "actions"], "notes": "The count is explicitly reported by the publisher, not inferred from the filename. Full training data remains outside this release.", "splits": ["training example subset"], "supervision": ["GUI actions"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset/tree/main", "gated": false, "hub_id": "mPLUG/UI_S1_dataset"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "d826463e2aa159f083dc74869bfabb8abb5b8ea4f2cdf4aeb7676172b18aaaa5", "supports": ["name", "kind", "relations", "licenses.data"], "url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "cdfbcbf23a3b1b4a609e1943e80dcf144ca1224a74ebf2308fde4fcb4326ab3c", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/mPLUG/UI_S1_dataset"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:57Z", "role": "primary", "sha256": "f668c890a5a87953a243715d0d47367c23f0c99b524d80f48bbdde8163f3e942", "supports": ["summary", "environments", "details.dataset_profile", "topics", "scope"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/UI-S1/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "d826463e2aa159f083dc74869bfabb8abb5b8ea4f2cdf4aeb7676172b18aaaa5", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md"}], "id": "url:16184830c0f4bdb146fa", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset"}], "name": "UI-S1 released training examples", "notes": "The release contains train_1000.jsonl and images.parquet. The official README explicitly describes 1,000 released examples and calls the release only part of the training dataset. Apache-2.0 is declared by the dataset card; this is not the complete training corpus.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md", "target_id": "arxiv:2509.11543", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A public example subset of the UI-S1 mobile training data, packaged as JSONL annotations and Parquet images.", "topics": ["reinforcement-learning", "data-generation", "mobile-navigation"], "updated_at": null, "url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo2-4B/tree/main", "gated": false, "hub_id": "Hcompany/Holo2-4B"}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "32232445026dc93113beb6b23d3e43da44d6288a67501f8f0023a78a43efb100", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo2-4B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "bb870c78f3075002ae42a7d1f392fda048fa31a58c865290b2a9b4a1137c6dd3", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo2-4B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "32232445026dc93113beb6b23d3e43da44d6288a67501f8f0023a78a43efb100", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo2-4B/raw/main/README.md"}], "id": "url:16a51d9e725ad5a5c629", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/Hcompany/Holo2-4B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo2-4B"}], "name": "Holo2 4B", "notes": "Two public safetensors shards are present and this 4B model card declares Apache-2.0. The same card assigns different terms to the 30B family member, so the 4B declaration is not generalized to the family. Scores and proprietary training data are not treated as independently verified or publicly released.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A Qwen3-VL-based checkpoint for navigation and multi-step GUI actions across web, desktop and mobile interfaces.", "topics": ["grounding", "planning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/Hcompany/Holo2-4B"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["target localization", "identification of impossible requests"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md", "population": "instruction-variant design", "qualifier": "exact", "unit": "instruction types", "value": 4}], "evaluation": "Publisher supplies original, action, description and negative instruction types; no common end-to-end success metric is asserted.", "evidence_urls": ["https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md", "https://huggingface.co/api/datasets/likaixin/ScreenSpot-v2-variants"], "focus": ["instruction-form robustness", "negative grounding requests"], "notes": "Reuses ScreenSpot-v2 imagery. Image-file count is not treated as a count of independent screenshots or tasks.", "observations": ["screenshots", "instructions"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/tree/main", "gated": false, "hub_id": "likaixin/ScreenSpot-v2-variants"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "7fb2bc9499d8a2b37b663c116af49d11bba493e6767d76a5305fad93257bebf2", "supports": ["name", "kind", "summary", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "dea98efa75e063eee2d6bfb5d46da94b41df3d779e2fb0bdaea7866534b5c469", "supports": ["identity", "release_status", "details", "environments"], "url": "https://huggingface.co/api/datasets/likaixin/ScreenSpot-v2-variants"}], "id": "url:17729e1c532ba7043404", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants"}], "name": "ScreenSpot-v2 instruction variants", "notes": "Public image files and desktop/mobile/web annotation JSON files are listed. This is a changed instruction set over an existing screenshot collection, not an independent new screenshot corpus. No data license or precise example count is declared in the card; no paper relationship is inferred from its generic project/PDF links.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Instruction variants over ScreenSpot-v2 screenshots that distinguish original commands, explicit actions, target descriptions and impossible operations.", "topics": ["grounding", "evaluation", "robustness"], "updated_at": null, "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/mPLUG/UI-S1-7B/tree/main", "gated": false, "hub_id": "mPLUG/UI-S1-7B"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "341abd3833131080682b27fe0b79e310d7cf9465602b77b8e399e45200552d61", "supports": ["name", "kind", "relations", "licenses.weights"], "url": "https://huggingface.co/mPLUG/UI-S1-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "d0c25fa2fad8372ffc8b51b63701c7e5548cca08a612c0081c4c317f9532894d", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/mPLUG/UI-S1-7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:57Z", "role": "primary", "sha256": "f668c890a5a87953a243715d0d47367c23f0c99b524d80f48bbdde8163f3e942", "supports": ["summary", "environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/UI-S1/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "341abd3833131080682b27fe0b79e310d7cf9465602b77b8e399e45200552d61", "supports": ["licenses.weights"], "url": "https://huggingface.co/mPLUG/UI-S1-7B/raw/main/README.md"}], "id": "url:1c4ee99a9a55ae71a7ac", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/mPLUG/UI-S1-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/mPLUG/UI-S1-7B"}], "name": "UI-S1 7B", "notes": "Four public safetensors shards are present. The card declares Apache-2.0 for the weights. The project publishes only a portion of its training data; no complete training corpus or independent reproduction is inferred.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/mPLUG/UI-S1-7B/raw/main/README.md", "target_id": "arxiv:2509.11543", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A mobile GUI checkpoint trained with semi-online reinforcement learning that uses offline trajectories to improve multi-turn action prediction.", "topics": ["reinforcement-learning", "mobile-navigation", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/mPLUG/UI-S1-7B"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["mobile UI actions", "clarification responses"], "counts": [], "evaluation": "Paper defines multi-path and noise/ambiguity evaluations; the public card identifies which files are full common-task sets and which are samples.", "evidence_urls": ["https://huggingface.co/datasets/xwk123/MobileBench-v2/raw/main/README.md", "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"], "focus": ["offline multi-path evaluation", "noisy interfaces", "ambiguous instructions"], "notes": "Release completeness differs by split. No full-benchmark sample total is inferred from the archive manifest.", "observations": ["mobile screenshots", "XML UI state", "instructions"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/xwk123/MobileBench-v2/tree/main", "gated": false, "hub_id": "xwk123/MobileBench-v2", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/xwk123/MobileBench-v2/raw/main/README.md"], "reason": "The same first-party card has an Apache-2.0 metadata header but explicitly assigns Apache-2.0 to code and CC-BY-NC-SA-4.0 to the dataset in its License section. Preserve both declarations and leave a single data license unresolved instead of displaying the permissive header as the data permission."}}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "13424be795c7b75b8c2195469c1252a7b3ec633e8bb31f1f6d7e44476611cbe9", "supports": ["name", "kind", "summary", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://huggingface.co/datasets/xwk123/MobileBench-v2/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "247b1766b8cbdb2943c4602d2667f9f06dec8b84f0d24457080d499177475f22", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/xwk123/MobileBench-v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["relations", "summary"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "356a8dca28fa3dc46bc05b13eaf04a3db273e63c06be4fd15d0e87e0343c0824", "supports": ["environments", "notes"], "url": "https://huggingface.co/api/datasets/xwk123/Mobile3M"}], "id": "url:1d5b85321b98765141fc", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xwk123/MobileBench-v2"}], "name": "Mobile-Bench-v2 released data", "notes": "The card distinguishes complete common-simple/common-complex JSON files from sampled noisy and ambiguous subsets; common tasks also require the separately cataloged Mobile3M dataset. Card metadata says Apache-2.0, while the prose explicitly assigns Apache-2.0 to code and CC-BY-NC-SA-4.0 to data. Do not apply the permissive header to dataset payloads, and do not describe every announced split as fully released.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50", "target_id": "arxiv:2505.11891", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Offline mobile-agent task data spanning common tasks and released noisy, contaminated and ambiguous-instruction subsets.", "topics": ["evaluation", "mobile-navigation", "robustness"], "updated_at": null, "url": "https://huggingface.co/datasets/xwk123/MobileBench-v2"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "e044e75ec1c2a74f2bee8c02b6efae05c399ec3537e34dd861dcf4a0a25dc182", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.browserless.io/"}], "id": "url:1e6b8682e0f1ddc3d36f", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browserless.io/"}], "name": "Browserless", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://www.browserless.io/", "target_id": "organization:browserless", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Hosted browser infrastructure for automation applications.", "topics": ["sandbox"], "updated_at": null, "url": "https://www.browserless.io/"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen2.5-VL-3B-Instruct", "gated": false, "hub_revision": "514c49e398fed1203e5c4192be9860131681299e", "license_name": "qwen-research", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "399895cd711d703ff4e016417ca14d13b76988c5625ff120e7083991eb0bd8f2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "c580cb5bd811285944337c4ebb461fa498bba09e6082b70e120ceb83e302fcce", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen2.5-VL-3B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "399895cd711d703ff4e016417ca14d13b76988c5625ff120e7083991eb0bd8f2", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md"}], "id": "url:23ad240757185c6f7eac", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B"}], "name": "Seg2Click Qwen2.5-VL 3B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference. The card names the Qwen Research license, unlike the Apache-2.0 declarations on the Qwen3.5 variants.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Seg2Click checkpoint based on Qwen2.5-VL-3B with a custom action-mask decoder.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "fa71ad6addd026ecd017059219c2457baf6b99c5c94f67350c0f9637c459c852", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://openai.com/index/introducing-chatgpt-atlas/"}], "id": "url:240a05c02f85d4da60dc", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/introducing-chatgpt-atlas/"}], "name": "ChatGPT Atlas", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://openai.com/index/introducing-chatgpt-atlas/", "target_id": "organization:openai", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "OpenAI browser product integrating ChatGPT and agent capabilities.", "topics": ["web-navigation"], "updated_at": null, "url": "https://openai.com/index/introducing-chatgpt-atlas/"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md", "population": "packaged tasks/videos", "qualifier": "exact", "unit": "demonstrations", "value": 350}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md", "population": "final packaged images", "qualifier": "exact", "unit": "evidence images", "value": 8388}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md", "population": "packaged action traces", "qualifier": "exact", "unit": "action records", "value": 11951}], "evidence_urls": ["https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md"], "formats": ["MP4", "PNG", "JSON", "JSONL", "commented key logs"], "modalities": ["screen video", "screenshots", "input events", "task text"], "notes": "The OSWorld subset was selected post hoc using model outcomes; no train/test split is supplied. Coordinate calibration, task success, and complete anonymization are not established by package-format checks.", "splits": [], "supervision": ["human workflow demonstrations", "skill induction"]}, "download_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/tree/main", "modalities": ["video", "image", "text", "input-events"], "sample_count": 350, "sample_unit": "task demonstrations / videos"}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:47:49Z", "role": "primary", "sha256": "ae14cfe91c8c9a79edf1f306273a2ee9d1790b1dc5be0cd2d89e93df09db42f0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T10:47:49Z", "role": "primary", "sha256": "ae14cfe91c8c9a79edf1f306273a2ee9d1790b1dc5be0cd2d89e93df09db42f0", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md"}], "id": "url:243ab00712ba19cc2089", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T10:47:49Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T10:47:49Z", "url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350"}], "name": "WorkDemo-350", "notes": "The card documents post hoc selection of the 286-task OSWorld subset based on model outcomes, uncalibrated coordinate annotations, and missing task-success acceptance records. It provides no train/test split or complete evaluator and does not claim complete anonymization. Review these limits before training or evaluation; the collection is an offline demonstration asset.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md", "target_id": "arxiv:2610.12984", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Desktop demonstrations with task descriptions, videos, input logs, and action-linked screenshot evidence for studying reusable agent skills.", "topics": ["recording", "workflow", "video"], "updated_at": null, "url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-wa/tree/main", "gated": false, "hub_id": "stanfordnlp/nnetnav-wa", "languages": [], "modalities": ["text"], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "3fe40a45aa2bf9d6f0cefffc65ea6382849c3543b44017200a4a8d43ab61e9eb", "supports": ["name", "kind"], "url": "https://huggingface.co/api/datasets/stanfordnlp/nnetnav-wa"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:34Z", "role": "primary", "sha256": "3095f7262b7d4ff916c6f58c55e6fc9902c08f77d09bf3951199945501148bb3", "supports": ["name", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "3fe40a45aa2bf9d6f0cefffc65ea6382849c3543b44017200a4a8d43ab61e9eb", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/stanfordnlp/nnetnav-wa"}], "id": "url:25a4d0ab29b60250863a", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:11:53Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T11:11:53Z", "url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-wa"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-wa"}], "name": "NNetNav WebArena data", "notes": "The Hub card supplies a license but no detailed schema. Dataset purpose is supported by the project's README; exact sample counts and annotation quality remain unverified.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md", "target_id": "github:murtyshikhar/nnetnav", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "NNetNav's WebArena-oriented browser-interaction data release, with separate training and test files.", "topics": ["data-generation", "web-navigation"], "updated_at": null, "url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-wa"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-4B", "gated": false, "hub_revision": "318dd331a8a2f60327dfd01e0ffc27998b460a3c", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "98409c81aeba5c9261075ab720204842465041f7df3aee833b28a8e5d4347c56", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "8c94e2115879f474e122cffafe4d14a4e898d832fe094d7f0fffcdba4a14f9d9", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-4B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "98409c81aeba5c9261075ab720204842465041f7df3aee833b28a8e5d4347c56", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md"}], "id": "url:29588aa5651b5839c045", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B"}], "name": "Seg2Click Qwen3.5 4B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Seg2Click checkpoint pairing a 4B Qwen3.5 backbone with pixel-level action-mask prediction.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/meituan/EvoCUA-8B-20260105/tree/main", "gated": false, "hub_id": "meituan/EvoCUA-8B-20260105", "languages": ["en", "zh"], "modalities": [], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "3fd1e50729e97746fec57c7d5da6f093d76fb9174e74c3f2c630fed5486fe995", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/meituan/EvoCUA-8B-20260105"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "3fd1e50729e97746fec57c7d5da6f093d76fb9174e74c3f2c630fed5486fe995", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/meituan/EvoCUA-8B-20260105"}], "id": "url:2a2682f86dcb8ebc9a80", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/meituan/EvoCUA-8B-20260105"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/meituan/EvoCUA-8B-20260105"}], "name": "EvoCUA 8B 20260105", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/meituan/EvoCUA-8B-20260105", "target_id": "github:meituan/evocua", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "EvoCUA computer-use agent checkpoint (20260105 release) with about 8.8B BF16 parameters in the Qwen3-VL architecture, tagged for OSWorld and for English and Chinese.", "topics": ["reinforcement-learning", "planning"], "updated_at": null, "url": "https://huggingface.co/meituan/EvoCUA-8B-20260105"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK", "population": "VideoCAD published collection", "qualifier": "more-than", "unit": "video recordings", "value": 41000}], "evidence_urls": ["https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK", "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md"], "formats": ["ZIP", "MP4", "JSON", "PNG"], "modalities": ["video", "UI action logs", "target CAD images"], "notes": "The reported video total is distinct from the Dataverse package-file count. The exact-version paper identifies the recorded interface as browser-based Onshape; operating-system compatibility is not established.", "splits": [], "supervision": ["timestamped UI actions", "target model images"]}, "dataverse": {"file_count": 261, "license": {"iconUri": "https://licensebuttons.net/p/zero/1.0/88x31.png", "languageCode": "en", "name": "CC0 1.0", "rightsIdentifier": "CC0-1.0", "rightsIdentifierScheme": "SPDX", "schemeUri": "https://spdx.org/licenses/", "uri": "http://creativecommons.org/publicdomain/zero/1.0"}, "persistent_url": "https://doi.org/10.7910/DVN/WX8PCK", "publication_date": "2025-05-15", "publisher": "Harvard Dataverse", "released_at": "2026-03-11T01:35:02Z", "version_minor_number": 0, "version_number": 5, "version_state": "RELEASED"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:20Z", "role": "primary", "sha256": "f67df0a4422a0484aaa2f93cd185b7000070d2a88d919d9ef995587f5e724007", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:53.536297+00:00", "role": "primary", "sha256": "cbb08803922b4b7736979ca0a5fe552b111fcd428fa9317401bc090b14bb4e54", "supports": ["summary", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:34.311949+00:00", "role": "primary", "sha256": "a303d95a811331cd6ee23b0340f1b9ef2dd996835bfbc66656a871ff94eead50", "supports": ["environments", "notes"], "url": "https://arxiv.org/html/2505.24838v2"}, {"method": "dataverse_api", "retrieved_at": "2026-10-17T02:03:20Z", "role": "primary", "sha256": "f67df0a4422a0484aaa2f93cd185b7000070d2a88d919d9ef995587f5e724007", "supports": ["details.dataverse", "licenses.data"], "url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK"}], "id": "url:2b5dead3172c0b70da7e", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:20Z", "licenses": {"code": null, "data": {"basis": "publisher Dataverse dataset-version license metadata; original declaration and URI retained in details.dataverse.license", "declared": "CC0 1.0", "retrieved_at": "2026-10-17T02:03:20Z", "url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/WX8PCK"}], "name": "VideoCAD data", "notes": "Dataverse reports a RELEASED version 5.0 with 261 unrestricted files and a CC0-1.0 dataset declaration. Archive count is not a video or trajectory count. The author repository documents the action and video formats, and Section 3 of the v2 paper identifies the recorded Onshape browser interface. No archives or training samples were downloaded; operating-system coverage is unspecified.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md", "target_id": "arxiv:2505.24838", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "CAD screen recordings paired with action logs and target model images for studying precise interface control and long-horizon visual imitation.", "topics": ["grounding", "imitation-learning", "video"], "updated_at": null, "url": "https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/WX8PCK"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md", "population": "paper-described benchmark coverage", "qualifier": "exact", "unit": "software applications", "value": 83}], "evidence_urls": ["https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md"], "formats": ["JSON annotations", "image files"], "modalities": ["screenshots", "bounding boxes", "UI labels", "action annotations"], "notes": "This is an offline evaluation dataset. Application coverage is not a sample count. The inspected file-layout documentation describes grounding annotations and images; it does not establish a common split or complete action-trajectory inventory.", "splits": [], "supervision": ["element grounding", "layout grounding", "action prediction"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/ServiceNow/ui-vision/tree/main", "gated": false, "hub_id": "ServiceNow/ui-vision", "languages": [], "modalities": ["image"], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "34f4ad5c9e7160b5be5aa8ae77e6e9f83a277aada41a272b82a2628d21159a8c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/ServiceNow/ui-vision"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:37Z", "role": "primary", "sha256": "295cf7f7500a18cea5c80eab1cdf4d7264f2f0e959b75efc5ba59083d93c9097", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "34f4ad5c9e7160b5be5aa8ae77e6e9f83a277aada41a272b82a2628d21159a8c", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/ServiceNow/ui-vision"}], "id": "url:2c2004baae4669f8169a", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/ServiceNow/ui-vision"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/ServiceNow/ui-vision"}], "name": "UI-Vision", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/ServiceNow/ui-vision", "target_id": "arxiv:2503.15661", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Desktop GUI data for interface understanding and interaction research.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/ServiceNow/ui-vision"}
{"aliases": ["url:1ff6b78a764bdcc6dd03", "HanXiao1999/UI-Genie-Agent-5k", "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-5k"], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md", "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md"], "formats": ["JSONL", "PNG"], "modalities": ["screenshots", "text", "actions"], "notes": "Public JSONL annotations and screenshots are present. MIT is declared in the dataset card. The release name, model-card links and annotation filename use different size labels; this review does not infer a reliable trajectory count from them.", "splits": [], "supervision": ["GUI action trajectories"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/tree/main", "gated": false, "hub_id": "HanXiao1999/UI-Genie-Agent-16k"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "a748fc7dbb53d5d751c5ee109e437c377e214b1d837762f8d09995349d86cec9", "supports": ["name", "kind", "relations", "licenses.data"], "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "548df402288350920a26bff2099544bf21835200ee58ff2c657c331fa6200317", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/HanXiao1999/UI-Genie-Agent-16k"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "e08569bd9d8a6ff44537d4fe69243ddfeb8a41ae5c9f87ef84af50a67c64c726", "supports": ["summary", "environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "a748fc7dbb53d5d751c5ee109e437c377e214b1d837762f8d09995349d86cec9", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:09Z", "role": "primary", "sha256": "66006a586e03be3eb1dfc29dc9c951241fd7e4204430e7db258da1037e17ac6e", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/HanXiao1999/UI-Genie-Agent-5k"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:10Z", "role": "primary", "sha256": "66006a586e03be3eb1dfc29dc9c951241fd7e4204430e7db258da1037e17ac6e", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/HanXiao1999/UI-Genie-Agent-16k"}], "id": "url:2e1927450da2183a5abe", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:34Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:01:34Z", "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k"}], "name": "UI-Genie trajectory data", "notes": "Public JSONL annotations and screenshots are present. MIT is declared in the dataset card. The release name, model-card links and annotation filename use different size labels; this review does not infer a reliable trajectory count from them.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md", "target_id": "arxiv:2505.21496", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Synthetic mobile GUI trajectory supervision and screenshots released with the UI-Genie agent-training pipeline.", "topics": ["data-generation", "mobile-navigation", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Atlas-Base-4B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "20c6679849cd81a871809aa7f96145855739c9170608dfa6ec3653e52bd795cb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-4B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "20c6679849cd81a871809aa7f96145855739c9170608dfa6ec3653e52bd795cb", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-4B"}], "id": "url:30c831a1198b9bc2a0f5", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B"}], "name": "OS-Atlas Base 4B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-4B", "target_id": "github:os-copilot/os-atlas", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "OS-Atlas base checkpoint with about 4.1B parameters, fine-tuned from InternVL2-4B and loaded in Transformers through custom InternVL chat model code.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": [], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md", "population": "Hub test split at revision b8eca829", "qualifier": "exact", "unit": "task-definition records", "value": 25}], "evaluation": "The published task schema contains rubrics and human_reference_operations. No automated scoring protocol or single numeric agent metric is established by the inspected card; execution uses separately configured Android tooling.", "evidence_urls": ["https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md", "https://huggingface.co/api/datasets/mikewang/mobile_eval_e", "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md", "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"], "focus": ["long-horizon multi-app Android task execution"], "notes": "Features are task_id, instruction, type, apps, scenario, rubrics and human-reference operation strings. A task record is not an action count or recorded screenshot trajectory. Public Parquet inventory was checked but no sample rows were downloaded; companion README limits the tested agent platform to Android.", "observations": ["task instructions", "app/scenario metadata"], "setting": "interactive"}, "disabled": false, "download_url": "https://huggingface.co/datasets/mikewang/mobile_eval_e/tree/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83", "gated": false, "hub_id": "mikewang/mobile_eval_e", "hub_revision": "b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md"], "reason": "No data-license declaration is present in the inspected task card; companion code MIT does not establish a dataset grant."}}, "repository_url": "https://github.com/X-PLUG/MobileAgent/tree/main/Mobile-Agent-E"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:15Z", "role": "primary", "sha256": "f8515fbce73ea0ef3d1d558d4567bb949244e6384e61c956ba3ddbc27294f0ae", "supports": ["name", "kind", "environments", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:29Z", "role": "primary", "sha256": "f9b10eada9e0a64e26f1a7d477768f218b807103199855d893b3ab07feddc755", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/mikewang/mobile_eval_e"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:15Z", "role": "primary", "sha256": "9fdf43753fbfac9868d8bec57fa5601db3e03ed5c30fb981f9da88f6b75ffcc5", "supports": ["summary", "details.benchmark_protocol", "licenses.data"], "url": "https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["summary", "topics", "details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "id": "url:342b081c3668a8bbe155", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T04:27:15Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/mikewang/mobile_eval_e"}], "name": "Mobile-Eval-E task definitions", "notes": "The public Hub manifest contains one test Parquet file. This resource is the benchmark’s task-definition package, not a bundled phone environment or agent. Card metadata supplies rubrics/reference-operation fields but no data-license declaration or verified automated scorer; the code repository’s MIT license is not transferred to these records.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md", "target_id": "arxiv:2501.11733", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Task definitions for Mobile-Agent-E’s long-horizon, multi-app Android evaluation, packaged as a 25-record test split with rubrics and human-reference operation strings.", "topics": ["evaluation", "long-horizon", "cross-app"], "updated_at": null, "url": "https://huggingface.co/datasets/mikewang/mobile_eval_e"}
{"aliases": [], "authors": ["Shaon Barman", "Sarah Chasins", "Ras Bodik", "Sumit Gulwani"], "details": {"publication": {"authors": ["Shaon Barman", "Sarah Chasins", "Ras Bodik", "Sumit Gulwani"], "date_basis": "Supplied month precision: 2016-11. year=2016, month=November", "date_parts": [2016, 11], "doi": "10.1145/2983990.2984020", "event_dates": null, "evidence_urls": ["https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/bibtex/", "https://2016.splashcon.org/track/splash-2016-oopsla", "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/", "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The original manuscript and MSR BibTeX print Sarah Chasins and Ras Bodik. The official conference profiles expand these to Sarah E. Chasins and Rastislav Bodík; the printed publication names are preserved. Page range, volume and exact publication day are not established by the retained citation sources.", "pages": null, "pdf_url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf", "publisher": null, "series": null, "title": "Ringer: Web Automation by Demonstration", "type": "conference-paper", "url": "https://doi.org/10.1145/2983990.2984020", "venue": "OOPSLA 2016", "volume": null}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:36Z", "role": "primary", "sha256": "633aa15e2e14346241de5385cc19d99f1fcdc9ef79fd06d27176dbfa8f38149a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/bibtex/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:36Z", "role": "primary", "sha256": "94c6b6bbd127126c9dd259bd346a7757d75366b4d805bdee2959c4c73226df74", "supports": ["details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://2016.splashcon.org/track/splash-2016-oopsla"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:24Z", "role": "primary", "sha256": "45f65483c3801dac142f4657c4cbe45e17b6a3e081497731e984c1e670c6e8f6", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:00Z", "role": "primary", "sha256": "bfbb69f57aaed39e3a210fae832e1fd98ed62ba1c01217b792238fe57e1a0ad1", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url"], "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:24Z", "role": "primary", "sha256": "45f65483c3801dac142f4657c4cbe45e17b6a3e081497731e984c1e670c6e8f6", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.title", "environments", "name", "scope", "topics"], "url": "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:36Z", "role": "primary", "sha256": "633aa15e2e14346241de5385cc19d99f1fcdc9ef79fd06d27176dbfa8f38149a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.type", "details.publication.venue", "kind"], "url": "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/bibtex/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:36Z", "role": "primary", "sha256": "94c6b6bbd127126c9dd259bd346a7757d75366b4d805bdee2959c4c73226df74", "supports": ["details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://2016.splashcon.org/track/splash-2016-oopsla"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:00Z", "role": "primary", "sha256": "bfbb69f57aaed39e3a210fae832e1fd98ed62ba1c01217b792238fe57e1a0ad1", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "summary"], "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf"}], "id": "url:347925d0f440d6396080", "identifiers": {"doi": "10.1145/2983990.2984020"}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/2983990.2984020"}, {"rel": "publication_pdf", "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf"}], "name": "Ringer: Web Automation by Demonstration", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The original manuscript and MSR BibTeX print Sarah Chasins and Ras Bodik. The official conference profiles expand these to Sarah E. Chasins and Rastislav Bodík; the printed publication names are preserved.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Ringer, which converts demonstrated webpage interactions into scripts combining actions, target selectors, and waiting conditions. It evaluates replay robustness on tasks recorded from real websites.", "topics": ["web-navigation", "recording", "robustness"], "updated_at": null, "url": "https://doi.org/10.1145/2983990.2984020"}
{"aliases": ["url:cab7bd78abb1faa469e7", "Reallm-Labs/InfiGUI-R1-3B", "https://huggingface.co/Reallm-Labs/InfiGUI-R1-3B"], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/tree/main", "gated": false, "hub_id": "InfiX-ai/InfiGUI-R1-3B"}, "environments": ["web", "desktop", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:41Z", "role": "primary", "sha256": "73496fb30838ba9ea291bb6d1e14004ac3318ac03a1ae7b5d6044e0dc14fd46f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:41Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/InfiX-ai/InfiGUI-R1-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["identity"], "url": "https://huggingface.co/api/models/Reallm-Labs/InfiGUI-R1-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "5fd1399c88bdf406e7689765ccf0599b3b7e0ddbed082fa92bffdd7b120bd41a", "supports": ["environments"], "url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:05:41Z", "role": "primary", "sha256": "73496fb30838ba9ea291bb6d1e14004ac3318ac03a1ae7b5d6044e0dc14fd46f", "supports": ["licenses.weights"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:15Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["aliases"], "url": "https://huggingface.co/api/models/Reallm-Labs/InfiGUI-R1-3B"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:16Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["aliases"], "url": "https://huggingface.co/api/models/InfiX-ai/InfiGUI-R1-3B"}], "id": "url:35b9879a351cc3b3d7d5", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:05:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:05:41Z", "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B"}], "name": "InfiGUI-R1 3B", "notes": "The official Hub API resolves the former Reallm-Labs model namespace to InfiX-ai. Two safetensors shards and model/tokenizer files are public and ungated. The card declares Apache-2.0 for this checkpoint; base-model terms remain separate. Its older loading example still uses the prior namespace.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md", "target_id": "arxiv:2504.14239", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL checkpoint trained with the Actor2Reasoner procedure to combine interface grounding with planning and recovery reasoning.", "topics": ["grounding", "reinforcement-learning", "self-reflection"], "updated_at": null, "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B"}
{"aliases": [], "authors": ["Toby Jia-Jun Li", "Amos Azaria", "Brad A. Myers"], "details": {"publication": {"authors": ["Toby Jia-Jun Li", "Amos Azaria", "Brad A. Myers"], "date_basis": "Supplied year precision: 2017. Publications: SUGILITE, CHI 2017 Only publication year is asserted; event dates are kept separate.", "date_parts": [2017], "doi": "10.1145/3025453.3025483", "event_dates": null, "evidence_urls": ["https://www.cs.cmu.edu/~NatProg/sugilite.html", "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Related to later APPINITE and PUMICE work but distinguished here by its 2017 three-author publication and base demonstration-automation contribution. Publication month/day are not independently identified by the retained sources.", "pages": "6038-6049", "pdf_url": "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf", "publisher": null, "series": null, "title": "SUGILITE: Creating Multimodal Smartphone Automation by Demonstration", "type": "conference-paper", "url": "https://doi.org/10.1145/3025453.3025483", "venue": "Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:22Z", "role": "primary", "sha256": "32ee05ff1a91ed4a22d4bc156a462872aa4e28f5c79abfbce5dbf3655f412033", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://www.cs.cmu.edu/~NatProg/sugilite.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:39Z", "role": "primary", "sha256": "cc23c19284f0a558c4b99b8b051abca8ed61e274e9e1f1fed1ae91487f4368ad", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.url"], "url": "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:39Z", "role": "primary", "sha256": "cc23c19284f0a558c4b99b8b051abca8ed61e274e9e1f1fed1ae91487f4368ad", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.url", "environments", "name", "scope", "summary", "topics"], "url": "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:22Z", "role": "primary", "sha256": "32ee05ff1a91ed4a22d4bc156a462872aa4e28f5c79abfbce5dbf3655f412033", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue", "kind"], "url": "https://www.cs.cmu.edu/~NatProg/sugilite.html"}], "id": "url:392a602c660ffa91be46", "identifiers": {"doi": "10.1145/3025453.3025483"}, "kind": "paper", "last_verified_at": "2026-10-17T09:54:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/3025453.3025483"}, {"rel": "publication_pdf", "url": "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf"}], "name": "SUGILITE: Creating Multimodal Smartphone Automation by Demonstration", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. Related to later APPINITE and PUMICE work but distinguished here by its 2017 three-author publication and base demonstration-automation contribution.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces SUGILITE, which combines spoken instructions, demonstrated Android actions, and UI hierarchies to generate reusable automation scripts. Its interface includes context checking and repairs for unfamiliar situations and changing app interfaces.", "topics": ["mobile-navigation", "recording", "human-in-the-loop", "cross-app"], "updated_at": null, "url": "https://doi.org/10.1145/3025453.3025483"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b/tree/main", "gated": false, "hub_id": "xTRam1/plan-and-act-planner-70b", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/xTRam1/plan-and-act-planner-70b/raw/main/README.md"], "reason": "The model card declares MIT in its metadata header (license: mit), beside a template comment asking the publisher to update it if needed. It is recorded as the publisher's declaration and does not override the Llama-3.3-70B-Instruct base model terms."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "4ef2d84d2b5d9d2972907e6135ea49f1ef088beaf6ef6042dc7d8fb4297a9a3a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.weights"], "url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "d2f757790ded81244e7c9b9e7350ec9fdc304609e387a95c7918db705bda535c", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/xTRam1/plan-and-act-planner-70b"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "feee31aa4d0801cb364578099a13e4f9cdcaa4eb5fbf4312b3fc3f8edd479a14", "supports": ["environments"], "url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/README.md"}], "id": "url:3932d69c6b6f9b2a7ba8", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b"}], "name": "Plan-and-Act Planner 70B", "notes": "The Hub manifest contains 17 safetensors shards and tokenizer/configuration files. Card metadata declares MIT and Llama-3.3-70B-Instruct as the base, but retains template comments asking the publisher to update those fields if needed. These are recorded as publisher declarations, without claiming they override upstream model terms.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b/raw/main/README.md", "target_id": "arxiv:2503.09572", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The planner component of Plan-and-Act, generating structured high-level plans that an executor turns into web actions.", "topics": ["planning", "web-navigation", "long-horizon"], "updated_at": null, "url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["browser interaction"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md", "population": "main V1/V2 suite (shared identities)", "qualifier": "exact", "unit": "task identities", "value": 609}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md", "population": "refusal suite", "qualifier": "exact", "unit": "tasks", "value": 111}], "evaluation": "Task-specific precomputed rubrics; refusal tasks are a separate evaluation population.", "evidence_urls": ["https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md"], "focus": ["live web task completion", "unsafe-request refusal"], "notes": "V2 refreshes task strings and rubrics while preserving V1 task IDs. Counts are task definitions, not independently verified runnable website states.", "observations": ["web pages"], "setting": "interactive"}, "disabled": false, "download_url": "https://huggingface.co/datasets/microsoft/WebTailBench/tree/main", "gated": false, "hub_id": "microsoft/WebTailBench", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md"], "reason": "The benchmark's Hub dataset card declares MIT in its metadata header (license: mit)."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "c4cfc361d1d47d7855269af4110483363964e23f9c65244a1182830eccccbb7f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "93469a20d333219dde89a225891a54ee716120d559eb8ac9f69a1a7ee84f57a3", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/microsoft/WebTailBench"}], "id": "url:3b6dfd52f98bfe76bc7f", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:42:32Z", "url": "https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/microsoft/WebTailBench"}], "name": "WebTailBench", "notes": "Public TSV files contain V1, V2 and refusal task definitions. The card declares MIT for the dataset. V1 and V2 share the same 609 task identities and must not be summed as distinct tasks. The May 2026 refresh supersedes calendar-bound task wording; live-site availability is not established by this metadata review.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md", "target_id": "arxiv:2511.19663", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Live-web task definitions and rubrics for evaluating computer-use agents, with preserved original tasks, a refreshed V2 task suite and a separate refusal set.", "topics": ["evaluation", "web-navigation", "safety-security"], "updated_at": null, "url": "https://huggingface.co/datasets/microsoft/WebTailBench"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/showlab/ShowUI-pi-data/raw/main/README.md"], "formats": ["Parquet", "LeRobot metadata"], "modalities": ["screenshots", "instructions", "actions"], "notes": "No sample or episode count is inferred from the size tag or package manifest. Main-branch files were inspected; the card’s v3.0 usage example was not executed.", "splits": [], "supervision": ["drag trajectories", "click episodes"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data/tree/main", "gated": false, "hub_id": "showlab/ShowUI-pi-data"}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "0e3a2a923865ae24d0de590143e9e32564b0e20db059579582d6181cdcf75bf1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "86fadf9a32185f018f6a8e49d52b39ee5c99c177075510361202dc25a214191f", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-pi-data"}], "id": "url:3dab53c6c38c9e7217ea", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data"}], "name": "ShowUI-π training data", "notes": "Public Parquet chunks and LeRobot metadata are present. The card names a v3.0 revision in its usage example; this review verifies the main manifest and does not assert that tag’s compatibility. No data license is declared. Hub size-category bands are not treated as episode counts.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data/raw/main/README.md", "target_id": "arxiv:2512.24965", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Drag-and-click episode data for the ShowUI-π model, spanning desktop editing, presentation, handwriting and other interface tasks.", "topics": ["grounding", "action-prediction", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/tree/main", "gated": false, "hub_id": "KDEGroup/UI-AGILE-Data"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:43:19Z", "role": "primary", "sha256": "ce35413bc1ef28338bc6fd9f132fa271a4f5e0bf23cecd7346f54adf0a964fd7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:43:19Z", "role": "primary", "sha256": "7782422de88830a497d9eb530bd2f0fb6283904cbb25dafcf2ff4d03358a377a", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/KDEGroup/UI-AGILE-Data"}], "id": "url:41abb1b7206689cf5057", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:43:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data"}], "name": "UI-AGILE-Data", "notes": "Public JSON, image archives and Parquet evaluation files are present. The card has no data-license declaration; the model’s MIT declaration is not inherited. The publisher describes fixes to AndroidControl action parameters, but those corrections were not independently reproduced or applied to upstream resources.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/raw/main/README.md", "target_id": "arxiv:2507.22025", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "UI-AGILE training annotations and screenshots together with converted GUI-grounding and AndroidControl evaluation files.", "topics": ["grounding", "reinforcement-learning", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data"}
{"aliases": [], "authors": [], "details": {"availability": "Ungated mixture/post-training JSON and image-archive manifests; some per-domain images and training code remain pending.", "dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md", "https://raw.githubusercontent.com/hkust-nlp/GUIMid/HEAD/README.md"], "formats": ["JSON", "tar.gz image archives"], "modalities": ["text", "images"], "notes": "Do not treat Hub size bands or the inconsistent introductory task count as a sample total. Some constituent-domain image packages remain forthcoming.", "splits": ["train configurations for named mid-training domains"], "supervision": ["mid-training instruction supervision", "GUI post-training trajectories"]}, "gated": false, "hub_id": "hkust-nlp/GUIMid", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md"], "reason": "Publisher dataset-card declaration for the released packaging; original source datasets retain their own terms."}}, "release_revision": "b6a360815c843b5733674191b6e6d7fa68b2ea8c"}, "environments": ["web", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:07Z", "role": "primary", "sha256": "1cdf702c819ecf370f4191c6c71fae7b579cfbaf766908b42156964a69eddacc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile", "licenses.data"], "url": "https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:07Z", "role": "primary", "sha256": "ededa2b3c8f21af8188152b028758c800db76cdbb8461ba82b8574cab375914e", "supports": ["identity", "release_status", "details.hub_id", "details.gated", "details.release_revision"], "url": "https://huggingface.co/api/datasets/hkust-nlp/GUIMid"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:19Z", "role": "primary", "sha256": "31b836da66c6907ff9a75ebc7f06da0db67e7d9fe9ca92d0c770f5f15f90a39e", "supports": ["identity"], "url": "https://api.github.com/repos/hkust-nlp/GUIMid"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:19Z", "role": "primary", "sha256": "d9f63a03088c37f7145725c7a182100d1593fceadd26a9bf817d434ca8d6297e", "supports": ["notes", "relations"], "url": "https://raw.githubusercontent.com/hkust-nlp/GUIMid/HEAD/README.md"}], "id": "url:422010da7a5c845a18f6", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T11:23:07Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for the released packaging; original source datasets retain their own terms.", "declared": "mit", "retrieved_at": "2026-10-17T11:23:07Z", "url": "https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/hkust-nlp/GUIMid"}], "name": "GUIMid training data", "notes": "The public manifest includes GUIMid.json, mavis.tar.gz, GUI_trajectory.json, traj.tar.gz and domain-specific JSON files. The card says image data for some individual domains is still forthcoming and its overview inconsistently says nine tasks while later enumerating seven vision-language plus four language-only domains; no aggregate task/sample count is inferred. Training code remains marked coming soon in the repository. MIT is the publisher dataset declaration, not a relicensing of all upstream content.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md", "target_id": "arxiv:2504.10127", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A released GUI-agent mid-training mixture and supporting domain datasets, together with a post-training trajectory file and image archives for web and mobile interaction.", "topics": ["imitation-learning", "planning", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/hkust-nlp/GUIMid"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/api/datasets/OS-Copilot/ScienceBoard-Traj", "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md"], "formats": ["ZIP"], "modalities": ["screenshots", "accessibility trees", "text"], "notes": "Observation modes follow publisher archive labels rather than inspected record schemas. No count or training-use permission is inferred.", "splits": [], "supervision": ["evaluation rollout logs"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/OS-Copilot/ScienceBoard-Traj/tree/main", "gated": false, "hub_id": "OS-Copilot/ScienceBoard-Traj", "languages": [], "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "0a97233bd9132fdddc6ddb26fd3cf9f2af06ae086429f367a61cfab0c16c4f44", "supports": ["name", "kind", "identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/OS-Copilot/ScienceBoard-Traj"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "4c3fe93aa740dbf4a279a246ddd836f2590a2a36f39b820abe166e38cb484577", "supports": ["summary", "environments", "relations", "topics", "scope"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "8ee436f67e708bb917992764bc60a017c9db2edb21bdb99356b35c69f73724d7", "supports": ["environments", "notes"], "url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "0a97233bd9132fdddc6ddb26fd3cf9f2af06ae086429f367a61cfab0c16c4f44", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OS-Copilot/ScienceBoard-Traj"}], "id": "url:4784f63cd7117f64a2a3", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/ScienceBoard-Traj"}], "name": "ScienceBoard evaluation trajectories", "notes": "The manifest exposes six ZIP archives but no README card or data-license declaration. The official project announces evaluation-result trajectories; this record does not label them as a training corpus or infer a trajectory count from archive names. Internal schemas and complete payload contents were not inspected.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md", "target_id": "arxiv:2505.19897", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Archives of ScienceBoard agent evaluation runs, separated by agent configuration and screenshot, accessibility-tree or set-of-marks observation mode.", "topics": ["evaluation", "science", "workflow"], "updated_at": null, "url": "https://huggingface.co/datasets/OS-Copilot/ScienceBoard-Traj"}
{"aliases": ["OpenCUA 72B Preview", "OpenCUA-72B-preview", "url:8c792416d297293422cf", "xlangai/OpenCUA-72B-preview", "https://huggingface.co/xlangai/OpenCUA-72B-preview"], "authors": [], "details": {"availability": "Public ungated checkpoint with 82 safetensors shards and model index.", "base_model": "Qwen/Qwen2.5-VL-72B-Instruct", "disabled": false, "download_url": "https://huggingface.co/xlangai/OpenCUA-72B/tree/main", "gated": false, "hub_id": "xlangai/OpenCUA-72B", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md"], "reason": "Publisher model card explicitly declares MIT for the released model; other code and training-data scopes are not generalized."}}, "release_revision": "3cc4af5337a1d0010f934284df3c482797266547"}, "environments": ["desktop", "linux", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:12Z", "role": "primary", "sha256": "ebe551af1008a2307d16cf68715fc1861f8c501c4981660dbe08f691cf73d787", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details", "licenses.weights"], "url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:12Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:35Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["identity", "notes"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B-preview"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:04Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["aliases"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B-preview"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:05Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["aliases"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B"}], "id": "url:4a0b1692589594745a3c", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T09:14:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher model card explicitly declares MIT for the released model; other code and training-data scopes are not generalized.", "declared": "mit", "retrieved_at": "2026-10-17T09:14:12Z", "url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/xlangai/OpenCUA-72B"}], "name": "OpenCUA 72B", "notes": "The former OpenCUA-72B-preview API redirects to the canonical xlangai/OpenCUA-72B repository. The public manifest lists 82 safetensors weight shards plus an index and custom model/tokenizer code. The card declares MIT for the released model and describes absolute coordinates after smart resizing. Weight files and remote code were not downloaded or executed.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md", "target_id": "github:xlang-ai/opencua", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md", "target_id": "arxiv:2508.09123", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-based computer-use checkpoint that combines screenshot history and reflective reasoning to predict executable desktop actions.", "topics": ["grounding", "planning", "action-prediction", "self-reflection"], "updated_at": null, "url": "https://huggingface.co/xlangai/OpenCUA-72B"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "7d67747fdfe89bcc69d2b60b41f4f2dccf758913ea6d491949cbeac20de372ef", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://hcompany.ai/newsroom/meet-holotab"}], "id": "url:4a47a6df845735984bf1", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://hcompany.ai/newsroom/meet-holotab"}], "name": "HoloTab", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://hcompany.ai/newsroom/meet-holotab", "target_id": "organization:h", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "H browser extension for executing tasks through website interfaces.", "topics": ["web-navigation"], "updated_at": null, "url": "https://hcompany.ai/newsroom/meet-holotab"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B/tree/main", "gated": false, "hub_id": "ByteDance-Seed/UI-TARS-1.5-7B", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "73b8964fb8ae2c25009dd4621549a3549d357c4c5c766ef7ed441acaef7efbc9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/ByteDance-Seed/UI-TARS-1.5-7B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "73b8964fb8ae2c25009dd4621549a3549d357c4c5c766ef7ed441acaef7efbc9", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/ByteDance-Seed/UI-TARS-1.5-7B"}], "id": "url:4d32ad3682ae2d71c178", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B"}], "name": "UI-TARS 1.5 7B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/ByteDance-Seed/UI-TARS-1.5-7B", "target_id": "github:bytedance/ui-tars", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "UI-TARS 1.5 GUI-agent checkpoint in the Qwen2.5-VL architecture with about 8.3B parameters stored in FP32, released under Apache-2.0 for English image-text-to-text use in Transformers.", "topics": ["grounding", "planning"], "updated_at": null, "url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "population": "RL training corpus", "qualifier": "exact", "unit": "prompts", "value": 12600}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "population": "RL training corpus", "qualifier": "exact", "unit": "screenshots", "value": 11205}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "population": "RL training corpus", "qualifier": "exact", "unit": "desktop applications", "value": 87}], "evidence_urls": ["https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"], "formats": ["Parquet"], "modalities": ["screenshots", "text", "coordinates"], "notes": "Publisher-reported counts cover a selection and reformatting of GroundCUA, not a new disjoint image collection. No independent image-overlap audit was performed.", "splits": ["train"], "supervision": ["instruction-to-click grounding", "target bounding boxes"]}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "34f0025f686d0e48e2baa5244056a4362018aaad77ac54abea7f1f6eb25c4276", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "34f0025f686d0e48e2baa5244056a4362018aaad77ac54abea7f1f6eb25c4276", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"}], "id": "url:4e71dc456c4112eed727", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL"}], "name": "GroundGUI RL Dataset", "notes": "The card declares MIT and attributes the source examples to GroundCUA. Its separate claim about UI-Vision overlap is not independently validated by this catalog.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Desktop grounding prompts selected from GroundCUA and formatted for click-based reinforcement learning with target-box rewards.", "topics": ["grounding", "reinforcement-learning", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-live/tree/main", "gated": false, "hub_id": "stanfordnlp/nnetnav-live", "languages": [], "modalities": ["text"], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "d8f8571dfc583d3a46d5b3a6a285e77f17057b36a8111d416d711513aa4ecf16", "supports": ["name", "kind"], "url": "https://huggingface.co/api/datasets/stanfordnlp/nnetnav-live"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:34Z", "role": "primary", "sha256": "3095f7262b7d4ff916c6f58c55e6fc9902c08f77d09bf3951199945501148bb3", "supports": ["name", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "d8f8571dfc583d3a46d5b3a6a285e77f17057b36a8111d416d711513aa4ecf16", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/stanfordnlp/nnetnav-live"}], "id": "url:4f47abde8cd28a1ec1e1", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:11:53Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T11:11:53Z", "url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-live"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-live"}], "name": "NNetNav Live data", "notes": "The Hub card supplies a license but no detailed schema. Dataset purpose is supported by the project's README; exact sample counts and annotation quality remain unverified.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md", "target_id": "github:murtyshikhar/nnetnav", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "Browser-interaction data released with NNetNav for its live-website setting, with separate training and test files.", "topics": ["data-generation", "web-navigation"], "updated_at": null, "url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-live"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md", "population": "recommended hflqf88888/GUIOdyssey release", "qualifier": "exact", "unit": "episodes", "value": 8334}], "evidence_urls": ["https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md", "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md", "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/introduction.md"], "formats": ["JSON annotations", "PNG screenshots"], "modalities": ["screenshots", "instructions", "actions", "bounding boxes", "step context"], "notes": "The repository gives conflicting episode totals (8,334 and 8,834). This profile uses 8,334 from the specifically recommended current dataset card. No independent archive recount was performed. Coordinates are normalized to 0–1000.", "splits": ["random", "task", "device", "app"], "supervision": ["cross-app trajectories", "action grounding"]}, "download_url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/tree/main", "sample_count": 8334, "sample_unit": "episodes reported by the current release card"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:39:52Z", "role": "primary", "sha256": "fd627ee0a6ee0596e0ee8147996b9a7a3436f754fc95ed63cf5f408d778b2831", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "4585b765d7c8b3e9bb6187d36d576edde454d4c0158a6aa8b3fdbedf34c89a15", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:39:52Z", "role": "primary", "sha256": "78c19ee712487af60df1913f11572958564e59a1a2cc663b1b30d971855d7a28", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/introduction.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T11:39:52Z", "role": "primary", "sha256": "fd627ee0a6ee0596e0ee8147996b9a7a3436f754fc95ed63cf5f408d778b2831", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md"}], "id": "url:511903ebfac45e6751b2", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:39:52Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T11:39:52Z", "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey"}], "name": "GUIOdyssey — current data release", "notes": "The project recommends this release over its earlier OpenGVLab/GUI-Odyssey upload. Episode count follows the current data card; conflicting repository prose is documented in the profile.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md", "target_id": "arxiv:2406.08451", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Cross-app Android interaction episodes with fine-grained step instructions, screen descriptions, action context, and target boxes.", "topics": ["mobile-navigation", "cross-app", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey"}
{"aliases": [], "authors": ["Tianlin Shi", "Andrej Karpathy", "Linxi Fan", "Jonathan Hernandez", "Percy Liang"], "details": {"publication": {"authors": ["Tianlin Shi", "Andrej Karpathy", "Linxi Fan", "Jonathan Hernandez", "Percy Liang"], "date_basis": "Publisher citation_publication_date metadata explicitly states 2017/07/17; this is separate from the conference interval.", "date_parts": [2017, 7, 17], "doi": null, "event_dates": "06--11 Aug (conference interval reported in the publisher BibTeX month field)", "evidence_urls": ["https://proceedings.mlr.press/v70/shi17a.html", "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Publisher citation uses Tianlin Shi and Linxi Fan; the PDF also prints the nicknames Tim and Jim. The ordered publisher author names are retained. DOI not supplied in the retained publisher citation. Publisher metadata supplies the publication date separately from the conference interval.", "pages": "3135-3144", "pdf_url": "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "World of Bits: An Open-Domain Platform for Web-Based Agents", "type": "conference-paper", "url": "https://proceedings.mlr.press/v70/shi17a.html", "venue": "Proceedings of the 34th International Conference on Machine Learning", "volume": "70"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "749d2a150e4c2bcde1b219a0ecb86c95546134a057892544e8495b5e45323c44", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.publisher", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v70/shi17a.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:39Z", "role": "primary", "sha256": "087008efb9d056df5fd9fbc724413d6a3f20cb14c7f1597d00bacac53d8342d6", "supports": ["details.publication.notes", "details.publication.pdf_url"], "url": "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "749d2a150e4c2bcde1b219a0ecb86c95546134a057892544e8495b5e45323c44", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume", "kind", "name", "scope", "summary", "topics"], "url": "https://proceedings.mlr.press/v70/shi17a.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:39Z", "role": "primary", "sha256": "087008efb9d056df5fd9fbc724413d6a3f20cb14c7f1597d00bacac53d8342d6", "supports": ["details.publication.notes", "details.publication.pdf_url", "environments"], "url": "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf"}], "id": "url:52a3261eb6a19b402e88", "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://proceedings.mlr.press/v70/shi17a.html"}, {"rel": "publication_pdf", "url": "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf"}], "name": "World of Bits: An Open-Domain Platform for Web-Based Agents", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. Publisher citation uses Tianlin Shi and Linxi Fan; the PDF also prints the nicknames Tim and Jim. The ordered publisher author names are retained.", "published_at": "2017-07-17", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces World of Bits, a web-task platform combining human demonstrations, keyboard and mouse actions, and cached HTTP traffic for repeatable task environments. It evaluates agents trained through behavioral cloning and reinforcement learning.", "topics": ["web-navigation", "reinforcement-learning", "evaluation", "data-generation"], "updated_at": null, "url": "https://proceedings.mlr.press/v70/shi17a.html"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/tree/main", "gated": false, "hub_id": "SenseLLM/SpiritSight-Agent-8B"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "916bde91ff00aba22ff13ed3d8b4ca679bc8c198632f1d3bc9469660287cb5d3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "cc2564a3b2011a06f435e16b95cbf1e22072fd39e76fba844657773541586b37", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/SenseLLM/SpiritSight-Agent-8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "916bde91ff00aba22ff13ed3d8b4ca679bc8c198632f1d3bc9469660287cb5d3", "supports": ["licenses.weights"], "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}], "id": "url:52b7de0e43b1afc9c226", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B"}], "name": "SpiritSight-Agent 8B base", "notes": "The Hub manifest exposes four safetensors shards under SpiritSight-Agent-8B-base. The card declares Apache-2.0 and separately links InternVL terms in its model table. This is the published base checkpoint; benchmark-specific fine-tuning is discussed separately. No model code was executed.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md", "target_id": "arxiv:2503.03196", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An InternVL2-based visual GUI checkpoint distributed for grounding and interface navigation, with the publisher recommending task-specific fine-tuning of the base release.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B"}
{"aliases": ["url:a778c3411858c38c0f39", "OpenMobile-2/OpenMobile-Data", "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data"], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md", "population": "SFT collection", "qualifier": "exact", "unit": "trajectories", "value": 11643}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md", "population": "SFT collection", "qualifier": "exact", "unit": "decision steps", "value": 118582}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md", "population": "RL collection", "qualifier": "exact", "unit": "executable tasks", "value": 2475}], "evidence_urls": ["https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md"], "formats": ["JSON", "JSONL", "PNG"], "modalities": ["screenshots", "text", "tool calls"], "notes": "These are publisher-reported totals at three distinct grains, not values to sum. Earlier screenshots may be collapsed; each SFT example keeps at most three images. The card and repository disagree on the singular/plural RL filename, so consumers should inspect the actual file manifest.", "splits": ["sft/train", "rl/train"], "supervision": ["next-action supervision", "verifiable task outcomes"]}}, "environments": ["android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "abacf9a10759a9e7a8eeec5a266c96ed83742bd9296a19be95d09c9717e18b7c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "485279038266d9454df75e27a80a0a0d71c24e90700126d2dcfa0da921b1273d", "supports": ["relations", "environments"], "url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:18Z", "role": "primary", "sha256": "14b1211ed709147f4022f311c17303ca70faaffa22be143617eb431d966eec1a", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/OpenMobile-2/OpenMobile-Data"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:20Z", "role": "primary", "sha256": "14b1211ed709147f4022f311c17303ca70faaffa22be143617eb431d966eec1a", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/OpenMobile-2/OpenMobile-Data-v2"}], "id": "url:546e6c95825d1318c98d", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2"}], "name": "OpenMobile-Data-v2", "notes": "The reviewed card contains no explicit license declaration, so data-license status remains unknown. The repository code license is not inherited. The current code README points to the v2 dataset; an older project-page URL was not counted as a separate asset.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/README.md", "target_id": "arxiv:2610.14157", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile-agent supervision from emulator and simulator trajectories, plus executable tasks with starting states and success checks for reinforcement learning.", "topics": ["mobile-navigation", "data-generation", "reinforcement-learning", "tool-integration"], "updated_at": null, "url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/neulab/UIX-Qwen2/tree/main", "gated": false, "hub_id": "neulab/UIX-Qwen2"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "91f90e329b76c2ded26a62c41f648a73e685116aa732c86c201592e46a941c44", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "19bd4d126f3ccb4cb3049125d384c76964bd3169c051369ac0c7b64cef9f48b4", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/neulab/UIX-Qwen2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "fe0c03fe73529a29788d559103c4aed48dca6ed9ba3bfd5f151f902fa8858b38", "supports": ["identity", "summary"], "url": "https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "91f90e329b76c2ded26a62c41f648a73e685116aa732c86c201592e46a941c44", "supports": ["licenses.weights"], "url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md"}], "id": "url:54997a3c67e54abdf319", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "odc-by", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/neulab/UIX-Qwen2"}], "name": "UIX-Qwen2", "notes": "Four public safetensors shards are present. The model card declares odc-by, a data-oriented license designation; it is preserved as a publisher declaration rather than replaced with the code or base-model license. No independent legal scope or evaluation result is asserted.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md", "target_id": "arxiv:2410.13824", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A MultiUI-trained visual-language checkpoint for text-rich webpage understanding and interface tasks.", "topics": ["screen-understanding", "grounding"], "updated_at": null, "url": "https://huggingface.co/neulab/UIX-Qwen2"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "ScreenRef image inventory", "qualifier": "exact", "unit": "distinct screenshots", "value": 585981}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "full recipe", "qualifier": "exact", "unit": "annotation rows", "value": 633369}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "refusal subset of full recipe", "qualifier": "exact", "unit": "annotation rows", "value": 22475}], "evidence_urls": ["https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"], "formats": ["Parquet"], "modalities": ["screenshots", "text", "coordinates"], "notes": "The publisher reports these inventories in the companion annotation card. The refusal subset is included in the full recipe. Screenshot access is gated; no image download or independent corpus recount was performed.", "splits": ["train"], "supervision": ["instruction-to-click grounding", "bounding boxes", "absent-target abstention"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef/tree/main", "gated": "auto", "hub_id": "PrentisAI/ScreenRef", "languages": ["en"], "license_name": "restricted-research-only", "modalities": ["image", "text"], "size_category_reported": ["100K<n<1M"]}, "environments": ["desktop", "web", "android", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "634b479c7d089e50735f6d4a25f78a63f4d0c2a8c1955015e7641b7afd61c317", "supports": ["name", "kind", "details.gated", "details.hub_id", "licenses.data"], "url": "https://huggingface.co/api/datasets/PrentisAI/ScreenRef"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "96a87222645c3f823e6f67e50708792e2917ddeb8eea5aacdf2de8c5bbff165f", "supports": ["summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile"], "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "634b479c7d089e50735f6d4a25f78a63f4d0c2a8c1955015e7641b7afd61c317", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/PrentisAI/ScreenRef"}], "id": "url:56bda04bd8d3a438deb3", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:51Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "other", "retrieved_at": "2026-10-17T01:48:51Z", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef"}], "name": "ScreenRef", "notes": "The public Hub API reports gated=auto and a Restricted Research-Only declaration. The companion card distinguishes annotations from screenshots and limits reuse. Access terms were inspected as public metadata; no access agreement was accepted or gated files downloaded.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI screenshots paired with localization instructions and targets, including absent-target examples; image access is gated under publisher terms.", "topics": ["grounding", "data-generation", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef"}
{"aliases": [], "authors": [], "details": {"availability": "Ungated Parquet annotation files for train, validation, test and test_unseen_os; videos external.", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "train split", "qualifier": "exact", "unit": "video annotation records", "value": 19725}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "validation split", "qualifier": "exact", "unit": "video annotation records", "value": 495}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "test split", "qualifier": "exact", "unit": "video annotation records", "value": 100}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "test_unseen_os split", "qualifier": "exact", "unit": "video annotation records", "value": 50}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "publisher overview of MONDAY", "qualifier": "approximately", "unit": "detected scenes", "value": 333000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "publisher overview of MONDAY", "qualifier": "approximately", "unit": "identified actions", "value": 313000}], "evidence_urls": ["https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "https://raw.githubusercontent.com/runamu/monday/HEAD/README.md"], "formats": ["Parquet"], "modalities": ["text", "timestamps", "bounding boxes", "action labels"], "notes": "Split counts are video-annotation rows; scene and action totals are rounded publisher reports and are not added to row counts. Source videos/images are not included in the Hub files.", "splits": ["train", "validation", "test", "test_unseen_os"], "supervision": ["video-derived scene and action annotations"]}, "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "https://raw.githubusercontent.com/runamu/monday/HEAD/README.md", "https://raw.githubusercontent.com/runamu/monday/HEAD/LICENSE"], "reason": "No dataset-license declaration in the inspected card; Apache-2.0 applies to companion code and is not transferred to YouTube content or this annotation package."}}, "release_revision": "d7bd929cf629bf0767f154cbdd984eea1eba01e9"}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "62a346fd0d328deb5b56a652c8581850b0c8120f5c77e53b33f24b9baf937adb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile", "licenses.data"], "url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "f55c319a70c8ff988049363dc8fc266a9f1b1a09b8ac7096ee573f962e8f9492", "supports": ["identity", "release_status", "details.release_revision"], "url": "https://huggingface.co/api/datasets/runamu/MONDAY"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "cb3d2329571a80b6edeab98c04f8b63c7f4e976929788d6a8283238b0b7fdadc", "supports": ["notes", "details.dataset_profile", "licenses.data"], "url": "https://raw.githubusercontent.com/runamu/monday/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:09Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["notes", "licenses.data"], "url": "https://raw.githubusercontent.com/runamu/monday/HEAD/LICENSE"}], "id": "url:570654ff67ecefa2b955", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T12:32:07Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/runamu/MONDAY"}], "name": "MONDAY annotations", "notes": "The ungated Hub publishes four Parquet splits containing video identifiers and annotations. Original videos and extracted screenshots must be obtained separately; the code README explains that consumers download source videos from YouTube. A Windows Mobile unseen-OS test split is documented but is not labeled desktop Windows here. The companion code has an Apache-2.0 file, while the data card declares no data license and the repository defers third-party materials to original terms. Training scripts, pretrained checkpoints and the full generation pipeline remain planned updates.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "target_id": "arxiv:2505.12632", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Video-derived mobile navigation annotations that pair scene boundaries and screen or element boxes with action labels for training and evaluating GUI agents.", "topics": ["data-generation", "mobile-navigation", "video", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/runamu/MONDAY"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "de16b4fd611982cfaa160cb0fbdc1b0f0559b90983e452964527704a6de4e72d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool"}], "id": "url:57f17b7d0c18b2d7c8bf", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool"}], "name": "Claude Computer Use", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool", "target_id": "organization:anthropic", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Claude tool capability for screenshot-based computer interaction.", "topics": ["planning", "grounding"], "updated_at": null, "url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool"}
{"aliases": [], "authors": [], "details": {"license_review": {"document": {"evidence_urls": ["https://raw.githubusercontent.com/huggingface/smolagents/12c1bc820eca50ace6f80a21d90426d41d74f845/docs/source/en/examples/web_browser.md", "https://raw.githubusercontent.com/huggingface/smolagents/12c1bc820eca50ace6f80a21d90426d41d74f845/LICENSE"], "reason": "The retained repository-root file contains Apache-2.0 terms, but the reviewed Markdown has no license notice and the retained sources do not explicitly establish that this license applies to the tutorial document. Document reuse terms remain unknown; this is not a claim that no license exists. Code, data, weights and hosted-service terms are not inferred."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:42:59.429802+00:00", "role": "primary", "sha256": "0bd3804794c37f263db5cba3429523721854e95fe1477127f9fb83d4d0510f50", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://raw.githubusercontent.com/huggingface/smolagents/12c1bc820eca50ace6f80a21d90426d41d74f845/docs/source/en/examples/web_browser.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:41:14.002361+00:00", "role": "primary", "sha256": "736281d4d96fc310050a09e830801f739729cc01217fef0be1fd45874e4c68f6", "supports": ["identity", "name", "summary", "environments", "topics", "scope", "notes"], "url": "https://huggingface.co/docs/smolagents/v1.26.0/examples/web_browser"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:41:17.450466+00:00", "role": "primary", "sha256": "cf08eff3ce4ca2fd85f85ba1eebca477277bca0cb2a54613e172741392061365", "supports": ["notes"], "url": "https://api.github.com/repos/huggingface/smolagents/git/ref/tags/v1.26.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:51:48.091044+00:00", "role": "primary", "sha256": "9def26af20ef4a6c4091e5873b53737940b360e6ff7db2d606add92d6586c8fb", "supports": ["notes"], "url": "https://huggingface.co/docs/smolagents/v1.8.0/examples/web_browser"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:43:02.729855+00:00", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.document"], "url": "https://raw.githubusercontent.com/huggingface/smolagents/12c1bc820eca50ace6f80a21d90426d41d74f845/LICENSE"}], "id": "url:5857c840372512cfe3fc", "identifiers": {}, "kind": "learning", "last_verified_at": "2026-10-20T13:42:59.429802+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/docs/smolagents/examples/web_browser"}], "name": "Web Browser Automation with Agents (smolagents tutorial)", "notes": "The reviewed v1.26.0 page matches all eight code blocks in its commit-pinned Markdown source. It uses InferenceClientModel and a vision model, unlike the retained v1.8.0 example. The shown setup starts a local visible Chrome session, and its installation command does not pin dependency versions. Runtime API compatibility, model-provider access and example outcomes were not tested. The inspected repository root contains Apache-2.0 terms, but explicit applicability to this Markdown tutorial was not established from retained sources; document reuse terms remain unknown. Model weights, visited websites and hosted-service terms are separate. No installation or notebook execution occurred.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Versioned tutorial building a vision-guided CodeAgent with Selenium/Helium browser tools, screenshot callbacks and example web-navigation tasks.", "topics": ["web-navigation", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/docs/smolagents/examples/web_browser"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md", "population": "collection described in the dataset card; not an independently counted release-row total", "qualifier": "more-than", "unit": "tasks", "value": 6000}], "evidence_urls": ["https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md", "https://huggingface.co/api/datasets/sunblaze-ucb/AgentSynth"], "formats": ["JSONL", "ZIP trajectory packages"], "modalities": ["screenshots", "task instructions", "LLM thoughts", "commands"], "notes": "Card maps osworld/train to all_tasks.jsonl and insta/train to task_sequences.jsonl. Six difficulty levels and trajectory screenshots/thoughts/commands are documented. ZIP file labels are not trajectory counts; configurations and composed-task levels are not assumed disjoint.", "splits": ["train (osworld configuration)", "train (insta configuration)"], "supervision": ["synthetic tasks", "agent trajectories"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/tree/9ec3f3139f9786e81aee14b7fff1eb487cc7e991", "gated": false, "hub_id": "sunblaze-ucb/AgentSynth", "hub_revision": "9ec3f3139f9786e81aee14b7fff1eb487cc7e991", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}, "repository_url": "https://github.com/sunblaze-ucb/AgentSynth"}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:17Z", "role": "primary", "sha256": "c1669d8e3caa27f8cd47c17b243b6a8d42807438d8a6a36074668adc57d33757", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:20Z", "role": "primary", "sha256": "0d3797d75dd4089ece8c8d4624f2a726024d25be255255d8770f78de21d75906", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/sunblaze-ucb/AgentSynth"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:12Z", "role": "primary", "sha256": "42e8b8dead7caea3f8b21e93a075e8c895dfc30a25e1470bf97a90940bbf0106", "supports": ["details.repository_url"], "url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/README.md"}], "id": "url:58a209f2b3b970916396", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:35:17Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:17Z", "url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth"}], "name": "AgentSynth tasks and trajectories", "notes": "Public Hub manifest contains JSONL task files and trajectory ZIP packages. Counts describe the publisher-reported task collection, not independently counted archive rows or numbers inferred from ZIP filenames.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md", "target_id": "arxiv:2506.14205", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Published AgentSynth task definitions and recorded GUI trajectories, with OSWorld and InSTA training configurations and packaged screenshots, thoughts and commands.", "topics": ["data-generation", "long-horizon", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/yifeizhou/pae-llava-7b/tree/main", "gated": false, "hub_id": "yifeizhou/pae-llava-7b"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "c0994582f81415c9b0d37e4e5018752e48e511178c4b3898c207a042992db845", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/yifeizhou/pae-llava-7b/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "1b76d2aebbe15954eff2b8859b884ec6504f8b67ed0b1045d8b3a83c0d132d3a", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/yifeizhou/pae-llava-7b"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "3e5f24b3175afeba8635dacc7b66eba1ce89fea49996a20281ff2df5a41f1e24", "supports": ["environments", "summary"], "url": "https://raw.githubusercontent.com/amazon-science/PAE/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "c0994582f81415c9b0d37e4e5018752e48e511178c4b3898c207a042992db845", "supports": ["licenses.weights"], "url": "https://huggingface.co/yifeizhou/pae-llava-7b/raw/main/README.md"}], "id": "url:5d61c1b5fc8dd650518d", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-nc-4.0", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/yifeizhou/pae-llava-7b/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/yifeizhou/pae-llava-7b"}], "name": "PAE LLaVA 7B", "notes": "The model is publicly ungated and has seven safetensors shards plus tokenizer and configuration files. Its card declares CC-BY-NC-4.0 for this model. This weight declaration does not settle the separate PAE dataset or code licensing questions.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/yifeizhou/pae-llava-7b/raw/main/README.md", "target_id": "arxiv:2412.13194", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A released LLaVA checkpoint for visual browser interaction from the PAE autonomous task-proposal and policy-learning system.", "topics": ["web-navigation", "reinforcement-learning"], "updated_at": null, "url": "https://huggingface.co/yifeizhou/pae-llava-7b"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["click target or bounding box"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md", "population": "test split", "qualifier": "exact", "unit": "examples", "value": 1639}], "evaluation": "Assess localization against annotated targets; agentbrowse, humanbrowse and calendars are distinct source buckets.", "evidence_urls": ["https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md"], "focus": ["instruction-conditioned web localization"], "notes": "The card reports more than 100 websites but does not provide a precise website total. Grounding accuracy does not establish task completion.", "observations": ["web screenshots", "instructions"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/Hcompany/WebClick/tree/main", "gated": false, "hub_id": "Hcompany/WebClick"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "958bbf0eb9f61afd6c576673fbe6f7d84e6c4b8a9d4d8c4fc0e1224b5d8da30b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "c9576957a3cb14785bf274691b50f90b04567a442ed60ee8c39f0f73108359f4", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Hcompany/WebClick"}], "id": "url:5e2bd1b3398974b79c9c", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Hcompany/WebClick"}], "name": "WebClick", "notes": "Four public test Parquet shards are present. The dataset card declares Apache-2.0. This is an offline localization test, not an end-to-end web-task success benchmark; reported screenshots and samples are not summed as different populations.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md", "target_id": "arxiv:2506.02865", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A web-interface localization benchmark with screenshot instructions and target boxes, including agent browsing, human browsing and calendar cases.", "topics": ["grounding", "evaluation", "web-navigation"], "updated_at": null, "url": "https://huggingface.co/datasets/Hcompany/WebClick"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/meituan/EvoCUA-32B-20260105/tree/main", "gated": false, "hub_id": "meituan/EvoCUA-32B-20260105", "languages": ["en", "zh"], "modalities": [], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "f42d9537a05f2562b57e5d508f95c3d29a9107d06574318c5715c7da5fb2a86d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/meituan/EvoCUA-32B-20260105"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "f42d9537a05f2562b57e5d508f95c3d29a9107d06574318c5715c7da5fb2a86d", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/meituan/EvoCUA-32B-20260105"}], "id": "url:60a95f34072aeb2771cc", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/meituan/EvoCUA-32B-20260105"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/meituan/EvoCUA-32B-20260105"}], "name": "EvoCUA 32B 20260105", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/meituan/EvoCUA-32B-20260105", "target_id": "github:meituan/evocua", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "Larger EvoCUA computer-use agent checkpoint (20260105 release) with about 33B BF16 parameters in the Qwen3-VL architecture, tagged for OSWorld and for English and Chinese.", "topics": ["reinforcement-learning", "planning"], "updated_at": null, "url": "https://huggingface.co/meituan/EvoCUA-32B-20260105"}
{"aliases": ["url:5f4cb613bda51b45d9c7", "inclusionAI/UI-Venus-2-9B", "https://huggingface.co/inclusionAI/UI-Venus-2-9B"], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/inclusionAI/UI-Venus-2-9b/tree/main", "gated": false, "hub_id": "inclusionAI/UI-Venus-2-9B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "0097be2c0e5c3f3c8aace10de835b0d7c1075531cc7823287c3f5488aebaea08", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "0097be2c0e5c3f3c8aace10de835b0d7c1075531cc7823287c3f5488aebaea08", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:02Z", "role": "primary", "sha256": "3ec7929329e8e1b09f31dee4a3e67128d68cddc95220a5944f4cd6102dae327e", "supports": ["aliases"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9B"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:03Z", "role": "primary", "sha256": "3ec7929329e8e1b09f31dee4a3e67128d68cddc95220a5944f4cd6102dae327e", "supports": ["aliases"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b"}], "id": "url:611df37e5f3caa6befd7", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/inclusionAI/UI-Venus-2-9b"}], "name": "UI-Venus 2 9B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b", "target_id": "github:inclusionai/ui-venus", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "UI-Venus 2 GUI-agent checkpoint fine-tuned from Qwen3.5-9B, with about 9.4B BF16 parameters in the Qwen3.5 multimodal architecture, loaded through Transformers for image-text-to-text.", "topics": ["grounding", "planning"], "updated_at": null, "url": "https://huggingface.co/inclusionAI/UI-Venus-2-9b"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/ritzzai/GUI-R1/tree/main", "gated": false, "hub_id": "ritzzai/GUI-R1"}, "environments": ["desktop", "windows", "linux", "macos", "android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "3eff86951de7c36dbd4af006ab3f1350213bdfea384494e59c8abc007793c979", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/ritzzai/GUI-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "94c7c78bede11dde727940f56592a04594c1c59438812dff47540bf53c1bed7d", "supports": ["environments"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["licenses.weights"], "url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md"}], "id": "url:61a5321db10fc8a312b9", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:30Z", "url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/ritzzai/GUI-R1"}], "name": "GUI-R1 checkpoint bundle", "notes": "The manifest contains two safetensors shards in GUI-R1-3B and four in GUI-R1-7B. Both are counted as one Hub resource rather than duplicate root URLs. The model card declares MIT; this does not automatically apply to source training data or upstream base models.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md", "target_id": "arxiv:2504.10458", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A single Hub release containing GUI-R1 3B and 7B action-model checkpoints trained with reinforcement learning across interface platforms.", "topics": ["grounding", "reinforcement-learning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/ritzzai/GUI-R1"}
{"aliases": [], "authors": ["Tsung-Hsiang Chang", "Tom Yeh", "Robert C. Miller"], "details": {"publication": {"authors": ["Tsung-Hsiang Chang", "Tom Yeh", "Robert C. Miller"], "date_basis": "Supplied month precision: 2010-04. Date Issued: April 2010", "date_parts": [2010, 4], "doi": "10.1145/1753326.1753555", "event_dates": null, "evidence_urls": ["https://dspace.mit.edu/entities/publication/6a6097ba-1d97-43cf-bc70-04b5af02efd5"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Distinct visual-testing contribution from the 2009 Sikuli paper. The institutional abstract and published-version citation supply the retained evidence. Exact publication day and operating-system coverage are not established in the accessible institutional abstract.", "pages": "1535-1544", "pdf_url": null, "publisher": null, "series": null, "title": "GUI Testing Using Computer Vision", "type": "conference-paper", "url": "https://doi.org/10.1145/1753326.1753555", "venue": "Proceedings of the 28th International Conference on Human Factors in Computing Systems (CHI '10)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbace77c54d1b4ed5dcf825dcb0f4525c32e33ac885e70312e8bb4a731a94e80", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://dspace.mit.edu/entities/publication/6a6097ba-1d97-43cf-bc70-04b5af02efd5"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbace77c54d1b4ed5dcf825dcb0f4525c32e33ac885e70312e8bb4a731a94e80", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://dspace.mit.edu/entities/publication/6a6097ba-1d97-43cf-bc70-04b5af02efd5"}], "id": "url:6262c757080f8838e62c", "identifiers": {"doi": "10.1145/1753326.1753555"}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/1753326.1753555"}], "name": "GUI Testing Using Computer Vision", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. This is a separate visual-testing contribution from the 2009 Sikuli paper, not another version of the same article. MIT PDF download returned HTTP 429; its complete institutional abstract and citation metadata remain the retained review source.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Presents visual GUI test scripts that use images to specify actions and expected screen feedback. Recording both input events and screenshots can generate tests from demonstrations, supporting unit, regression, and test-driven workflows.", "topics": ["evaluation", "grounding", "automation"], "updated_at": null, "url": "https://doi.org/10.1145/1753326.1753555"}
{"aliases": [], "authors": ["James Allen", "Nathanael Chambers", "George Ferguson", "Lucian Galescu", "Hyuckchul Jung", "Mary Swift", "William Taysom"], "details": {"publication": {"authors": ["James Allen", "Nathanael Chambers", "George Ferguson", "Lucian Galescu", "Hyuckchul Jung", "Mary Swift", "William Taysom"], "date_basis": "Supplied year precision: 2007. Papers section: AAAI-2007", "date_parts": [2007], "doi": null, "event_dates": null, "evidence_urls": ["https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf", "https://www.cs.rochester.edu/research/cisd/projects/plow/"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The ordered names follow the original published paper, without expanding initials or replacing names from secondary indexes. DOI and volume number are not supplied in the retained primary sources. No exact publication month/day is established.", "pages": "1514-1519", "pdf_url": "https://cdn.aaai.org/AAAI/2007/AAAI07-240.pdf", "publisher": null, "series": null, "title": "PLOW: A Collaborative Task Learning Agent", "type": "conference-paper", "url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf", "venue": "Twenty-Second AAAI National Conference on Artificial Intelligence", "volume": null}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:26Z", "role": "primary", "sha256": "795d275c9f85180561fd101652221d30fb98daca2c2a00faf2b57c397bdfd68d", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.url"], "url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:24Z", "role": "primary", "sha256": "c0b3571efd5f2e137e42b38107de29aca6668819bec84109912d514867d85b8c", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://www.cs.rochester.edu/research/cisd/projects/plow/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:26Z", "role": "primary", "sha256": "795d275c9f85180561fd101652221d30fb98daca2c2a00faf2b57c397bdfd68d", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.url", "environments", "name", "scope", "summary", "topics"], "url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:24Z", "role": "primary", "sha256": "c0b3571efd5f2e137e42b38107de29aca6668819bec84109912d514867d85b8c", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue", "environments", "kind"], "url": "https://www.cs.rochester.edu/research/cisd/projects/plow/"}], "id": "url:628b5251757bb3888e89", "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf"}, {"rel": "publication_pdf", "url": "https://cdn.aaai.org/AAAI/2007/AAAI07-240.pdf"}], "name": "PLOW: A Collaborative Task Learning Agent", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The ordered names follow the original published paper, without expanding initials or replacing names from secondary indexes.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes PLOW, which learns executable web-task models from a session combining demonstrations, natural-language explanations, and dialogue. Browser instrumentation connects user actions with DOM structure to support task generalization.", "topics": ["web-navigation", "human-in-the-loop", "planning"], "updated_at": null, "url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/raw/main/README.md", "population": "Linux, macOS, Windows, Android, iOS, and Web", "qualifier": "exact", "unit": "platform categories", "value": 6}], "evidence_urls": ["https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/raw/main/README.md"], "formats": ["JSONL annotations", "JSON manifest", "split image archives"], "modalities": ["screenshots", "text", "actions"], "notes": "The six categories include Web, so they are not six operating systems. Grounding and trajectory annotations follow different collection processes. No sample total or train/test division is inferred from archive-part counts or Hub size tags.", "splits": [], "supervision": ["GUI grounding", "GUI action trajectories"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/tree/main", "gated": false, "hub_id": "OpenGVLab/ScaleCUA-Data", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "dca4b89176c7653e83545363e4868bd2d5f0bf618df007b68a7e25407262f1ac", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/OpenGVLab/ScaleCUA-Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:37Z", "role": "primary", "sha256": "8c545d66cc7199a9258bd91ed48930691d0d99555bc30ee457fae9bf656fd18d", "supports": ["details.dataset_profile", "relations"], "url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "dca4b89176c7653e83545363e4868bd2d5f0bf618df007b68a7e25407262f1ac", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OpenGVLab/ScaleCUA-Data"}], "id": "url:631e74223e198ef40202", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data"}], "name": "ScaleCUA Data", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/raw/main/README.md", "target_id": "arxiv:2509.15221", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Cross-platform training data released with ScaleCUA, with grounding and navigation annotations for Ubuntu, Windows, macOS, Android, iPhone and web screenshots.", "topics": ["data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "bb658d64714578f8ce5d78b014474781f865e42ccc078be3b61539bc3780b235", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://scribehow.com/"}], "id": "url:63ef8c7e455951a20281", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://scribehow.com/"}], "name": "Scribe", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "adjacent", "summary": "Workflow recording and documentation product adjacent to agent data collection.", "topics": ["recording", "workflow"], "updated_at": null, "url": "https://scribehow.com/"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md", "https://huggingface.co/api/datasets/cckevinn/GUI-Actor-Data", "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/data/data_config.yaml"], "formats": ["JSON annotations", "ZIP image packages", "multipart image archives"], "modalities": ["GUI screenshots", "bounding boxes", "grounding annotations"], "notes": "Public Hub revision 5a0b35fb includes six bbox JSON files and associated image archives. The training configuration names six input paths with sampling_strategy=all; those input paths are not six disjoint datasets or a verified example count. The card’s source-list wording and mismatched hyperlinks remain unresolved, so its approximately 1M screenshots/10M elements headline is not adopted as a release inventory. No data license is established by the card, and the model repository license is not transferred to these data.", "splits": ["training"], "supervision": ["GUI grounding with bounding-box supervision"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/tree/main", "gated": false, "hub_id": "cckevinn/GUI-Actor-Data"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "e42bbcf8d466c977f9b2150feb73d8cbda2ff907c279630ad4730a8b1f5d665c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "details.dataset_profile"], "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "d6271cff879b1a2a70490cf3dd95765bef44ecfca531d8ef3545f101673b4423", "supports": ["identity", "release_status", "details", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/cckevinn/GUI-Actor-Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dafdebf2fa1e0a5db871addd103e50cb55625f0108c757708029e885920ee4e", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:12Z", "role": "primary", "sha256": "3521c849b29d7fc88430a8e73a8b2aeb2e613d7fa16ca054dd5bf4ec011b36ca", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/data/data_config.yaml"}], "id": "url:66096ad4057a3d3726f5", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data"}], "name": "GUI-Actor data", "notes": "Public JSON annotations and image archives are present. The card has no data-license declaration. Its source list contains mismatched hyperlinks and says six datasets while naming fewer distinct entries; no source-derivation relationships or aggregate count are imported from that list.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md", "target_id": "arxiv:2506.03143", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A processed screenshot and bounding-box collection released for training GUI-Actor visual grounding models.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "df8615d0687cba27e5558ea3f7ecb1823871e3804522603dd62b52dcf127ac89", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.askui.com/"}], "id": "url:69d6afe38a9a128ba9f7", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.askui.com/"}], "name": "AskUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://www.askui.com/", "target_id": "organization:askui", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Agentic test-automation platform running plain-language test cases from pixels rather than selectors on desktop apps, Citrix, in-car head units and mobile devices, deployable on-premise or air-gapped.", "topics": ["workflow"], "updated_at": null, "url": "https://www.askui.com/"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/wanlilll/WeaveBench/tree/main", "gated": false, "hub_id": "wanlilll/WeaveBench", "languages": ["en"], "modalities": [], "size_category_reported": ["n<1K"]}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "e65e5793fe5b251ef7793a0618f4e8139fe5c92c2cf8093d0cc6142e8ece9d02", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/wanlilll/WeaveBench"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "e65e5793fe5b251ef7793a0618f4e8139fe5c92c2cf8093d0cc6142e8ece9d02", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/wanlilll/WeaveBench"}], "id": "url:6a212bc12f82c824f6b9", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:27Z", "url": "https://huggingface.co/datasets/wanlilll/WeaveBench"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/wanlilll/WeaveBench"}], "name": "WeaveBench", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Computer-use benchmark mixing graphical interaction and code execution.", "topics": ["evaluation", "tool-integration"], "updated_at": null, "url": "https://huggingface.co/datasets/wanlilll/WeaveBench"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/CUA-Gym/tree/main", "gated": false, "hub_id": "xlangai/CUA-Gym", "languages": ["en"], "modalities": ["tabular", "text"], "size_category_reported": ["10K<n<100K"]}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "43b8ebde3b971e1b8b61d789c786ead59b45a2490bc6ed19e204bafc03a31f44", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/xlangai/CUA-Gym"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:25Z", "role": "primary", "sha256": "962cd9162bea83f7adfedb3da50129bcf4203ae65176bdf53f0af6788b9603ef", "supports": ["relations"], "url": "https://huggingface.co/datasets/xlangai/CUA-Gym/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "43b8ebde3b971e1b8b61d789c786ead59b45a2490bc6ed19e204bafc03a31f44", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xlangai/CUA-Gym"}], "id": "url:6aa1aa58a7a888395189", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/xlangai/CUA-Gym"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/CUA-Gym"}], "name": "CUA-Gym", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/xlangai/CUA-Gym/raw/main/README.md", "target_id": "arxiv:2605.25624", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Computer-use environment and task resources for agent research.", "topics": ["evaluation", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/xlangai/CUA-Gym"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "single-step portion", "qualifier": "approximately", "unit": "action instructions", "value": 67000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "multi-step portion", "qualifier": "approximately", "unit": "action instructions", "value": 15000}], "evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"], "formats": ["JSON", "Parquet"], "modalities": ["screenshots", "instructions", "actions"], "notes": "Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.", "splits": ["train", "test"], "supervision": ["single-step actions", "multi-step navigation"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/yiye2023/GUIAct/tree/main", "gated": false, "hub_id": "yiye2023/GUIAct", "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "https://huggingface.co/datasets/yiye2023/GUIAct/raw/main/README.md"], "reason": "Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict."}}}, "environments": ["web", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "624ccdd6e6c5b29b588a4289d1d94c1a1c0803868bd586175f2c9ebc3bfdd0b6", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/yiye2023/GUIAct"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "98b45ea81164d1e1a1dd82255207053b15cd6c69d922a1c5cf3387ce604d4b74", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yiye2023/GUIAct/raw/main/README.md"}], "id": "url:6cd150ec23361b4d2e84", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yiye2023/GUIAct"}], "name": "GUIAct", "notes": "Public JSON annotations and Parquet image packages are present. The project README declares CC-BY-4.0 for datasets, while this Hub card declares Apache-2.0; these conflicting declarations are retained without asserting a single confirmed data license. The project code’s MIT declaration is separate.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Web and Android GUI action-instruction data, with single-step, multi-step and smartphone partitions.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/yiye2023/GUIAct"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/tree/main", "gated": false, "hub_id": "HanXiao1999/UI-Genie-Agent-7B"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "3d4e2deda3e8ecc2e6bbba732c01a84dee2443bb7132ef1aeae444f001d23d69", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "20e47378595ed007766a135c37fe6709d8853a780ecd5ef56792f0b88c05aeff", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/HanXiao1999/UI-Genie-Agent-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "3d4e2deda3e8ecc2e6bbba732c01a84dee2443bb7132ef1aeae444f001d23d69", "supports": ["licenses.weights"], "url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/raw/main/README.md"}], "id": "url:6d959b58bd4a2e98e417", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B"}], "name": "UI-Genie Agent 7B", "notes": "Four public safetensors shards are present and the card declares MIT. Its training-data metadata/link uses UI-Genie-Agent-5k while prose labels it 16k; no precise training-set relationship or count is asserted across those inconsistent references. Larger announced models are not counted as released.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/raw/main/README.md", "target_id": "arxiv:2505.21496", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL mobile action model trained using existing and reward-guided synthetic trajectories from the UI-Genie framework.", "topics": ["mobile-navigation", "imitation-learning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/LangAGI-Lab/WebShepherd_3B/tree/main", "gated": false, "hub_id": "LangAGI-Lab/WebShepherd_3B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "6f28e86a99b286a575c5e7f8d24d6901655971e9a091493950b59948f4b9ad9b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/LangAGI-Lab/WebShepherd_3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:05Z", "role": "primary", "sha256": "647d9e7bdad5fe2c03a8cfcd34d93f3b2bf7c1553b517a4a843bc0113fa47456", "supports": ["identity", "summary", "relations"], "url": "https://huggingface.co/api/collections/LangAGI-Lab/web-shepherd-advancing-prms-for-reinforcing-web-agents-682b4f4ad607fc27c4dc49e8"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "07431058d8b9a71b80e75b7a74a40a8346c1503eec240628459cc4d4e2f0802f", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:49:45Z", "role": "primary", "sha256": "252ed290ec84ffd8fa686fb72df68afe7bbd8c19b80538c0b318db60cd107d09", "supports": ["details"], "url": "https://huggingface.co/LangAGI-Lab/WebShepherd_3B/raw/main/config.json"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "6f28e86a99b286a575c5e7f8d24d6901655971e9a091493950b59948f4b9ad9b", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/LangAGI-Lab/WebShepherd_3B"}], "id": "url:72f7fe0211c6551b945f", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:45:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/LangAGI-Lab/WebShepherd_3B"}], "name": "WebShepherd 3B", "notes": "The official project-linked collection identifies this public 3B checkpoint. The README uses a separate unavailable WebShepherd/web-shepherd-base URL; redirect equivalence is not asserted. Two safetensors shards and a Qwen2 causal-language-model configuration are present. No model card or weight license is published in the inspected manifest. The separately announced multimodal 3B model is not assumed released.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/collections/LangAGI-Lab/web-shepherd-advancing-prms-for-reinforcing-web-agents-682b4f4ad607fc27c4dc49e8", "target_id": "arxiv:2505.15277", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A public text-based process-reward checkpoint from the Web-Shepherd collection for scoring web-agent decisions against task checklists.", "topics": ["reward-models", "web-navigation"], "updated_at": null, "url": "https://huggingface.co/LangAGI-Lab/WebShepherd_3B"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-0.8B", "gated": false, "hub_revision": "c36a096cb0cb3320099eb8fde82608cadf2248ee", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "8130a68ede5391bd158ce2cc49c36c5f8c8ebf22b897db6f68e27babbbc66965", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "6bd2608de2715e3d76a0c9e9b8fd7540372bfd2f9864db0445d8e548c9afb518", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-0.8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "8130a68ede5391bd158ce2cc49c36c5f8c8ebf22b897db6f68e27babbbc66965", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md"}], "id": "url:7405463ba56f9a41e79a", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B"}], "name": "Seg2Click Qwen3.5 0.8B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Compact Seg2Click checkpoint with a merged Qwen3.5 backbone and an action-mask decoder.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "14f573af0c4ea3e5922b8831b7f03724936963e293d7c0aaa3fccbb71dbb0cfa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://openai.com/index/computer-using-agent/"}], "id": "url:74eeb7c9fb20674fe6e2", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/computer-using-agent/"}], "name": "OpenAI CUA", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://openai.com/index/computer-using-agent/", "target_id": "organization:openai", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Computer-using model introduced with the Operator research preview.", "topics": ["planning", "grounding"], "updated_at": null, "url": "https://openai.com/index/computer-using-agent/"}
{"aliases": [], "authors": [], "details": {"artifact_type": "Ubuntu VMware image", "disabled": false, "download_url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env/tree/main", "gated": false, "hub_id": "OS-Copilot/ScienceBoard-Env", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/OS-Copilot/ScienceBoard-Env/raw/main/README.md", "https://huggingface.co/api/models/OS-Copilot/ScienceBoard-Env"], "reason": "This resource is a prebuilt Ubuntu VMware environment archive (VM.zip), not neural model weights. MIT Hub metadata in a model namespace cannot establish a weights license or relicense the guest OS and bundled applications."}}}, "environments": ["linux", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "8ee436f67e708bb917992764bc60a017c9db2edb21bdb99356b35c69f73724d7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.weights"], "url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "9ffaf62b54a610afe186b3576b55fdd8287289e0cdfb88371bd9849a616f8460", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/OS-Copilot/ScienceBoard-Env"}], "id": "url:75917f2c2b2802aafd3e", "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env"}], "name": "ScienceBoard VMware environment", "notes": "The public Hub manifest contains VM.zip. This is an environment image, not neural model weights despite its Hub namespace. The card declares MIT metadata for the hosted release, but that does not relicense Ubuntu or included applications. The archive was not downloaded or booted; VMware’s own terms are separate.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env/raw/main/README.md", "target_id": "arxiv:2505.19897", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A prebuilt Ubuntu virtual-machine archive containing the software environment used for ScienceBoard scientific computer-use tasks.", "topics": ["sandbox", "science", "evaluation"], "updated_at": null, "url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/osunlp/WebJudge-7B/tree/main", "gated": false, "hub_id": "osunlp/WebJudge-7B"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "425f11651cdae09997c1b038cdc9c091778c35a51b8de88140fc6dc935b8bb64", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/osunlp/WebJudge-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "2eebc1f909263be946057306b41d733e29f0f32337dae02da4e989b5477cf31f", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/osunlp/WebJudge-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "425f11651cdae09997c1b038cdc9c091778c35a51b8de88140fc6dc935b8bb64", "supports": ["licenses.weights"], "url": "https://huggingface.co/osunlp/WebJudge-7B/raw/main/README.md"}], "id": "url:772161f06c2b3d4978b5", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/osunlp/WebJudge-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/osunlp/WebJudge-7B"}], "name": "WebJudge 7B", "notes": "Four public safetensors shards are present. Apache-2.0 is declared for the weights. This is an evaluator checkpoint rather than a navigation policy; human-agreement rates are not treated as task-success scores.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/osunlp/WebJudge-7B/raw/main/README.md", "target_id": "arxiv:2504.01382", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A visual trajectory judge that retains selected intermediate screenshots to evaluate whether a web agent completed its task.", "topics": ["evaluation", "reward-models", "web-navigation"], "updated_at": null, "url": "https://huggingface.co/osunlp/WebJudge-7B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md", "population": "default/train in the dataset card metadata", "qualifier": "exact", "unit": "training rows", "value": 701864}], "evidence_urls": ["https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"], "formats": [], "modalities": ["screenshots", "instructions", "functionality labels", "bounding boxes"], "notes": "Rows are not unique screenshots. The card documents image-downscaling conventions that affect coordinates; preserve the device and original image-size fields.", "splits": ["train"], "supervision": ["element grounding", "functionality captioning"]}, "download_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/tree/main", "modalities": ["image", "text", "bounding-box"], "sample_count": 701864, "sample_unit": "training rows reported in the dataset card"}, "environments": ["web", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "c73b46acc2ff034a443fb8dbde327de1888535267199c9fb1819adf66898f1d7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "c73b46acc2ff034a443fb8dbde327de1888535267199c9fb1819adf66898f1d7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"}], "id": "url:79287058342289aa5d45", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T10:42:13Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T10:42:13Z", "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k"}], "name": "AutoGUI v1 training data", "notes": "The row count is not a unique-screen count. The data card documents image scaling conventions that affect coordinate interpretation.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md", "target_id": "arxiv:2502.01977", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screenshot-based grounding and captioning samples with UI functionality descriptions, target coordinates, and bounding boxes.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k"}
{"aliases": ["url:f4bcd696942ee04dcc42", "bevaya/ScreenSpot", "https://huggingface.co/datasets/bevaya/ScreenSpot"], "authors": [], "details": {"benchmark_protocol": {"actions": ["predicted click point; a predicted box is reduced to its center"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md", "population": "Hub test split at 0be08781; not unique screenshots", "qualifier": "exact", "unit": "instruction–target examples", "value": 1272}], "evaluation": "The SeeClick evaluator scores whether the predicted point falls inside the target box, with platform and text/icon breakdowns. The Hub packaging uses xyxy boxes, whereas the original evaluator reads xywh and converts it; adapt the loader to the chosen release.", "evidence_urls": ["https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md", "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot", "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/pretrain/screenspot_test.py"], "focus": ["single-step GUI element grounding"], "notes": "Offline, zero-shot evaluation across mobile, desktop, and web. The card explicitly excludes training use in its intended-use description. Public API reports ungated Parquet files and canonical Hub id bevaya/ScreenSpot; the catalog identity is preserved. Apache-2.0 is the Hub data declaration. No runtime or underlying screenshot rights audit was performed.", "observations": ["screenshot", "English instruction"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot/tree/main", "gated": false, "hub_id": "bevaya/ScreenSpot", "languages": ["en"], "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "9e2d2a2c076032225911ae1032b548eafa6331ea81a2779632610f0f1c82dfd9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:10:00Z", "role": "primary", "sha256": "5ab2240c064502f75a6726311e10e7724ab7055347697f883b15a6ee8e7d00a6", "supports": ["details.benchmark_protocol"], "url": "https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "eb72113eefd4868114f8a8105eba2d3385da9034a5b3a4860661e03ae7dcae15", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/pretrain/screenspot_test.py"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "9e2d2a2c076032225911ae1032b548eafa6331ea81a2779632610f0f1c82dfd9", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:56Z", "role": "primary", "sha256": "af0c4babeff84158ff196b54bf47242923bbf502d2805a1dd4eb52dba44be89c", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/bevaya/ScreenSpot"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:56Z", "role": "primary", "sha256": "af0c4babeff84158ff196b54bf47242923bbf502d2805a1dd4eb52dba44be89c", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot"}], "id": "url:7abb418096432f56c685", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot"}], "name": "ScreenSpot", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI grounding test set of 1,272 screenshot instructions from iOS, Android, macOS, Windows and web interfaces, each labeled with the target element's bounding box.", "topics": ["grounding", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "463db02b607e7215a4421440c4d274ee3546bd9eefc50ed9e10f64d27d7f8f38", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.browseract.com/"}], "id": "url:7ae2670859e3796e8607", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browseract.com/"}], "name": "BrowserAct", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://www.browseract.com/", "target_id": "organization:browseract", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Managed web scraping platform with reusable browser interaction workflows.", "topics": ["workflow"], "updated_at": null, "url": "https://www.browseract.com/"}
{"aliases": [], "authors": [], "details": {"architecture": "LlamaForCausalLM", "availability": "Ungated complete weight/config/tokenizer manifest.", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/META_LLAMA_3_COMMUNITY_LICENSE_AGREEMENT.txt", "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/README.md"], "reason": "Weight license text and card declaration are retained separately from framework code licensing."}}, "parameter_label": "8B", "release_revision": "09856c28ec324cf3fe4c513bdc4437a75b78185f", "training_data_note": "Publisher describes a 24K curated subset of WebLINX interactions and dialogue; not a newly counted independent dataset."}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "1a445c7894c65c9a7c4247dba446c206c13ac2ecdc129ce93a0de943f2984a18", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.weights"], "url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "69fe17825db0d5f32ccaa84b67fabfa10018e1410da0c70ec0ea9509703feba2", "supports": ["identity", "release_status", "details.release_revision", "details.architecture"], "url": "https://huggingface.co/api/models/McGill-NLP/Llama-3-8B-Web"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "6bcca341999ab474e2973116128177b77f97cd64028eeace224035358ce6474f", "supports": ["licenses.weights"], "url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/META_LLAMA_3_COMMUNITY_LICENSE_AGREEMENT.txt"}], "id": "url:7c3e8c4ba52428584079", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T12:34:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Weight license text and card declaration are retained separately from framework code licensing.", "declared": "Meta Llama 3 Community License Agreement", "retrieved_at": "2026-10-17T12:36:55Z", "url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/META_LLAMA_3_COMMUNITY_LICENSE_AGREEMENT.txt"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web"}], "name": "Llama-3-8B-Web", "notes": "The ungated Hub manifest includes model.safetensors, pytorch_model.bin, config and tokenizer files; the two weight formats represent the same checkpoint and are not counted separately. The card describes an approximately 24K-example curated WebLINX training subset and evaluates dialogue-guided web navigation. Predictions still require parsing and execution in an external browser environment. The retained weight license is the Meta Llama 3 Community License Agreement, not WebLlama code MIT. No weights were downloaded, model used, or agreement accepted.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/README.md", "target_id": "github:mcgill-nlp/weblinx", "type": "trained_on"}, {"evidence_url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/README.md", "target_id": "arxiv:2402.05930", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Llama 3 8B instruction-model checkpoint fine-tuned on WebLINX web interactions and dialogue, predicting browser actions from a textual state representation.", "topics": ["web-navigation", "imitation-learning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train/tree/main", "gated": false, "hub_id": "LZXzju/UI-R1-3B-Train"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "640cb5908abefb770ea6bb8d12f22e9eca622224fa73e9f5c67a3a95f4f7b8e5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "482d1304ad53736b3e410694cfa427067ff8b45267374964e4a8ff3833afbbd3", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/LZXzju/UI-R1-3B-Train"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f55ac6a0c50d298d46c5f7083a0879959fd7cd3c1d91496af295ebf2afcc13a3", "supports": ["environments", "summary"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md"}], "id": "url:7c568d88a472f914d316", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train"}], "name": "UI-R1 3B training data", "notes": "The public manifest includes train.json and a train_imgs directory. The card has no dataset-license declaration, and the model’s MIT metadata is not inherited. The official repository identifies this as v1 training data drawn from AndroidControl and ScreenSpot; later enlarged training mixtures are distinct.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train/raw/main/README.md", "target_id": "arxiv:2503.21620", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The released mobile training subset for the original UI-R1 3B model, with action annotations and individual screenshot files.", "topics": ["data-generation", "grounding", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Genesis-7B-AC"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "62de4360ad62165bd4dc37d5897d6f5f88813956dff33bac63a24ded50cca118", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "c35278027cd0e36e2fbbba4861031eca2a52090f20b40fb5aa680b067bb13dac", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Genesis-7B-AC"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "62de4360ad62165bd4dc37d5897d6f5f88813956dff33bac63a24ded50cca118", "supports": ["licenses.weights"], "url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/raw/main/README.md"}], "id": "url:80858580c1009eba0e98", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC"}], "name": "OS-Genesis 7B AC", "notes": "Four public safetensors shards are present. The card declares Apache-2.0. This is the AC task-specific checkpoint; AndroidWorld and WebArena variants are separate releases and their results are not transferred to it.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/raw/main/README.md", "target_id": "arxiv:2412.19723", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The AndroidControl-oriented Qwen2-VL checkpoint trained with mobile trajectories generated through OS-Genesis reverse task synthesis.", "topics": ["mobile-navigation", "action-prediction", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-9B", "gated": false, "hub_revision": "76b7e9f7ce97675b7d4e1ea6d88a6adc9ced2886", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "feeb30bda7f86e10d87bdc1b0148d0632aba6df1676931a3d6c65cbdf197df3e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "f215590b7a7b510085f7f99a8c236816604d4c11a8f5ae544b4ee75771dd4f82", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-9B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "feeb30bda7f86e10d87bdc1b0148d0632aba6df1676931a3d6c65cbdf197df3e", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md"}], "id": "url:826f3753905c5c43bf46", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B"}], "name": "Seg2Click Qwen3.5 9B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Seg2Click checkpoint pairing a 9B Qwen3.5 backbone with a decoder for valid GUI click regions.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Atlas-Base-7B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "a3355c4f884490efe2dc31ad393b869349c35130cc37370dafa19f4f0f5016b7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-7B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "a3355c4f884490efe2dc31ad393b869349c35130cc37370dafa19f4f0f5016b7", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-7B"}], "id": "url:82a30f45d9e855f423c7", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B"}], "name": "OS-Atlas Base 7B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-7B", "target_id": "github:os-copilot/os-atlas", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "OS-Atlas base checkpoint fine-tuned from Qwen2-VL-7B-Instruct, using the standard Qwen2-VL architecture in Transformers with about 8.3B BF16 parameters.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["PyAutoGUI script: pointer movement/click/drag, key presses, hotkeys, and text entry"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md", "population": "train split, HF revision 5e4bcb26", "qualifier": "exact", "unit": "task examples", "value": 6788}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md", "population": "test split, HF revision 5e4bcb26", "qualifier": "exact", "unit": "task examples", "value": 2020}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md", "population": "val split, HF revision 5e4bcb26", "qualifier": "exact", "unit": "task examples", "value": 991}], "evaluation": "The bundled evaluator reports weighted sequence match and action match, comparing action types and applying coordinate, key, and writing penalties. Target element-box metadata is evaluation-only. The retained scorer has a click-penalty defect: predicted coordinates are read from gold_line, so score reproduction requires explicit scorer-version review.", "evidence_urls": ["https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md", "https://huggingface.co/api/datasets/Writer/omniact", "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/eval.py"], "focus": ["offline desktop/web action-script generation"], "notes": "This released split table totals 9,799 examples; it is not silently substituted for any paper-level collection size. Public files include train/test/val JSON indexes, task/image archives, and eval.py. Evaluation is offline script comparison, not execution success on a live desktop. MIT is the Hub data declaration; original app/content rights are not resolved by it.", "observations": ["screenshot", "natural-language task"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/Writer/omniact/tree/main", "gated": false, "hub_id": "Writer/omniact", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "2db0a5dd2b983dded3b90a8ecd66cadbb39074f03778a339822418531661b3e3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/Writer/omniact"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:10:00Z", "role": "primary", "sha256": "7b54ad557f89d621fe39c3d2fe9f02164a2bfcb7fef0db610d964586d3673693", "supports": ["details.benchmark_protocol"], "url": "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:06Z", "role": "primary", "sha256": "a786de678af46df5e2b6b1a2a69d2ee239fe8fd5b63663765d84d39e21fc2365", "supports": ["details.benchmark_protocol"], "url": "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/eval.py"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "2db0a5dd2b983dded3b90a8ecd66cadbb39074f03778a339822418531661b3e3", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Writer/omniact"}], "id": "url:849a86beaee006c6b147", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:27Z", "url": "https://huggingface.co/datasets/Writer/omniact"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Writer/omniact"}], "name": "OmniAct", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/Writer/omniact", "target_id": "arxiv:2402.17553", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Dataset and benchmark for generating computer interaction programs.", "topics": ["evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/Writer/omniact"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md", "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md"], "formats": ["JSONL"], "modalities": ["text"], "notes": "The public manifest contains reward_data.jsonl but no separately packaged image archive. The card declares MIT. The 517k identifier is not converted into a verified sample count, and complete image packaging is not inferred from the JSONL filename.", "splits": [], "supervision": ["trajectory rewards"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/tree/main", "gated": false, "hub_id": "HanXiao1999/UI-Genie-RM-517k"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "dddb70b2b6986525f3980cbe50c3241d17f840a81fda9715b01ac3b25dd35d63", "supports": ["name", "kind", "relations", "licenses.data"], "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "ab4bf30b6188959f825f2747ca711b0f190dddfe1658319c881ca9c73be9ecc9", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/HanXiao1999/UI-Genie-RM-517k"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "e08569bd9d8a6ff44537d4fe69243ddfeb8a41ae5c9f87ef84af50a67c64c726", "supports": ["summary", "environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "dddb70b2b6986525f3980cbe50c3241d17f840a81fda9715b01ac3b25dd35d63", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md"}], "id": "url:84e5aeeffaf561baa966", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:34Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:01:34Z", "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k"}], "name": "UI-Genie reward data", "notes": "The public manifest contains reward_data.jsonl but no separately packaged image archive. The card declares MIT. The 517k identifier is not converted into a verified sample count, and complete image packaging is not inferred from the JSONL filename.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md", "target_id": "arxiv:2505.21496", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile GUI reward annotations released for training UI-Genie’s trajectory-assessment model.", "topics": ["data-generation", "mobile-navigation", "reward-models"], "updated_at": null, "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "e7062cfb4c0948d3a2274843f9f493f34cb40dfd634371408086df6141a0e036", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://ai.google.dev/gemini-api/docs/computer-use"}], "id": "url:879d4e3e75e83582bde6", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://ai.google.dev/gemini-api/docs/computer-use"}], "name": "Gemini Computer Use", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://ai.google.dev/gemini-api/docs/computer-use", "target_id": "organization:google", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Gemini capability for generating interface actions from screenshots.", "topics": ["planning", "grounding"], "updated_at": null, "url": "https://ai.google.dev/gemini-api/docs/computer-use"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "41ec857dacdb106ad99abbe1872f40a295056522ffaa1b1714b51dffb370f450", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.agentql.com/"}], "id": "url:8896d44370711136fd54", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.agentql.com/"}], "name": "AgentQL", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Natural-language querying and interaction tools for web interfaces.", "topics": ["grounding", "web-navigation"], "updated_at": null, "url": "https://www.agentql.com/"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md", "population": "train split", "qualifier": "exact", "unit": "rows", "value": 21988}], "evidence_urls": ["https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md"], "formats": ["Parquet", "JSON", "TAR.GZ"], "modalities": ["screenshots", "text", "bounding boxes", "points"], "notes": "Row count is the card dataset_info value; one row may contain multiple instructions. The card summarizes screenshots as approximately 22K. No data license is declared.", "splits": ["train"], "supervision": ["element grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/showlab/ShowUI-web/tree/main", "gated": false, "hub_id": "showlab/ShowUI-web"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "08586aa1ded1f1a1f984fb55697fdb1007d86c7fdf4988a2dc1fcf3fdfd49a47", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "12412480f1a3957069d9dcdc53abce4ad3e34be7dac7ca6fe1726d04764cadfd", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-web"}], "id": "url:8933977bb0a57940ed23", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/showlab/ShowUI-web"}], "name": "ShowUI-web", "notes": "Public Parquet annotations, metadata JSON and images.tar.gz are present. The card refers to image.tar.gz in its extraction command, while the actual archive is plural. No data-license declaration is present; the ShowUI code or model licenses are not inherited.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md", "target_id": "arxiv:2411.17465", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Web-interface grounding examples with instructions, element boxes and points, paired with a separately packaged screenshot archive.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/showlab/ShowUI-web"}
{"aliases": [], "authors": [], "details": {"availability": "Ungated annotation files and multipart image-archive manifest.", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md", "population": "TongUI-collected GUI-Net-1M data, excluding added external mixtures", "qualifier": "approximately", "unit": "training trajectories", "value": 1000000}], "evidence_urls": ["https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md", "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md"], "formats": ["JSON annotations", "multipart image archives"], "modalities": ["text", "screenshots"], "notes": "The source reports an approximate trajectory total; it does not establish per-platform counts or a complete test-split inventory.", "splits": ["training collection"], "supervision": ["GUI trajectory supervision from multimodal tutorials"]}, "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md"], "reason": "Publisher declaration for this dataset packaging; original web/tutorial materials retain applicable rights."}}, "release_revision": "60d1e644fcff7d3b62093baec03fe7efeb0d98a1"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "f30a18d1a533b8f71f436dae2ee7369f8971422bd76527e069499c5bba911912", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "314a77eb0ba43bedffbe6869601e41aea41993eaca59d86ba346652c853802d2", "supports": ["identity", "release_status", "details.release_revision", "details.dataset_profile.formats"], "url": "https://huggingface.co/api/datasets/Bofeee5675/GUI-Net-1M"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "b51acb70c95d079ccbe1de0a685ad7c8a4b3b5bed11e6162e6f7c2f60d78805e", "supports": ["summary", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "c2df5c3a6dd576ae53f0dc0f72ecab07d93ed034b04336abd2072a6f0d812c64", "supports": ["licenses.data", "details.dataset_profile.formats", "notes"], "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md"}], "id": "url:8c9aaafb918aaf8b4e92", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": {"basis": "Publisher declaration for this dataset packaging; original web/tutorial materials retain applicable rights.", "declared": "apache-2.0", "retrieved_at": "2026-10-17T12:32:07Z", "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M"}], "name": "GUI-Net-1M", "notes": "The ungated manifest lists JSON annotations and multipart image archives. Training documentation describes roughly one million collected training trajectories from Wikihow, Baidu Experience and GUI videos; this is distinct from larger mixtures that add external datasets. Platform-specific distribution and exact row counts were not established from this release metadata. Apache-2.0 is the publisher dataset declaration, without a grant of additional rights to upstream tutorial content. No image archive or example record was downloaded.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md", "target_id": "arxiv:2504.12679", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A TongUI-collected training corpus from multimodal web tutorials, distributed as annotation files and split screenshot archives for GUI trajectory supervision.", "topics": ["data-generation", "grounding", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/raw/main/README.md", "population": "Hub split named test", "qualifier": "exact", "unit": "rows", "value": 9460}], "evidence_urls": ["https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/raw/main/README.md", "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"], "formats": ["Parquet"], "modalities": ["screenshots", "text", "action history"], "notes": "Preserves the publisher’s split label despite the collection’s training purpose. Rows are not individual preference pairs because rejected is a list. Data license is unknown.", "splits": ["test"], "supervision": ["preferred and rejected continuations", "task checklists"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/tree/main", "gated": false, "hub_id": "LangAGI-Lab/WebPRMCollection_preference_pair"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "9c091c5fc625ce53f53f8262f5cee8d08194e4fd1b8eeed543ff1c3723c823fd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "28a1d5d12e84fc38515bd3c5031ae0cf315c5f8772b90d982d926b98f2857563", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/LangAGI-Lab/WebPRMCollection_preference_pair"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "07431058d8b9a71b80e75b7a74a40a8346c1503eec240628459cc4d4e2f0802f", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"}], "id": "url:8dfc84eb4016204fe70a", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair"}], "name": "WebPRM Collection preference data", "notes": "Ten public Parquet shards are present. The card names its split test and reports 9,460 rows; the project’s 40K-plus preference-annotation claim is a different grain and is not substituted for this row count. A row can contain multiple rejected alternatives. No data-license declaration is present.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/raw/main/README.md", "target_id": "arxiv:2505.15277", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Checklist-annotated web decision records with page observations, prior actions, one preferred continuation and rejected alternatives for process-reward learning.", "topics": ["web-navigation", "reward-models", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/openbmb/AgentCPM-GUI/tree/main", "gated": false, "hub_id": "openbmb/AgentCPM-GUI"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "9016a68e833a7717c922d9363a6d6c56cc7cf6a7087543f5694507668fe9ab75", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "4d76c8cc19f1ff4c850ba6cc566d50bbd2355e10921a2214e38c921cedfeed4c", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/openbmb/AgentCPM-GUI"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "9016a68e833a7717c922d9363a6d6c56cc7cf6a7087543f5694507668fe9ab75", "supports": ["licenses.weights"], "url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md"}], "id": "url:8ef3985dce697e5feea9", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/openbmb/AgentCPM-GUI"}], "name": "AgentCPM-GUI 8B checkpoint", "notes": "The Hub provides four safetensors shards and associated model files without gating. Weight metadata declares Apache-2.0, while the prose license section refers specifically to code; these scopes are kept distinct. The card describes Android-app operation, and the CAGUI evaluation data remains a separate existing catalog asset.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md", "target_id": "arxiv:2506.01391", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An eight-billion-parameter MiniCPM-V-based model that consumes smartphone screenshots and predicts compact actions for Chinese and English Android app tasks.", "topics": ["mobile-navigation", "grounding", "reinforcement-learning"], "updated_at": null, "url": "https://huggingface.co/openbmb/AgentCPM-GUI"}
{"aliases": [], "authors": ["Toby Jia-Jun Li", "Oriana Riva"], "details": {"publication": {"authors": ["Toby Jia-Jun Li", "Oriana Riva"], "date_basis": "Supplied month precision: 2018-06. Publication line: MobiSys 2018, June 2018", "date_parts": [2018, 6], "doi": "10.1145/3210240.3210339", "event_dates": null, "evidence_urls": ["https://www.microsoft.com/en-us/research/publication/kite-building-conversational-bots-mobile-apps/", "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Final proceedings page range and volume are not established in the retained first-party citation. The paper states a 14-page extent, not a page range. Exact publication day is unknown.", "pages": null, "pdf_url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf", "publisher": null, "series": null, "title": "Kite: Building Conversational Bots from Mobile Apps", "type": "conference-paper", "url": "https://doi.org/10.1145/3210240.3210339", "venue": "The 16th Annual International Conference on Mobile Systems, Applications, and Services", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:26Z", "role": "primary", "sha256": "e2f5403d77f563ec6a88458e32d057559a8f8ba4d2a80304e70c2c793fb60bb5", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes"], "url": "https://www.microsoft.com/en-us/research/publication/kite-building-conversational-bots-mobile-apps/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:27Z", "role": "primary", "sha256": "bfb67283f49dad7974d97e3b3da9d9ff24744d18e83c4d25cbe4f12ecf807071", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:27Z", "role": "primary", "sha256": "bfb67283f49dad7974d97e3b3da9d9ff24744d18e83c4d25cbe4f12ecf807071", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:26Z", "role": "primary", "sha256": "e2f5403d77f563ec6a88458e32d057559a8f8ba4d2a80304e70c2c793fb60bb5", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes"], "url": "https://www.microsoft.com/en-us/research/publication/kite-building-conversational-bots-mobile-apps/"}], "id": "url:901ca855640457dfe2f1", "identifiers": {"doi": "10.1145/3210240.3210339"}, "kind": "paper", "last_verified_at": "2026-10-17T09:59:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/3210240.3210339"}, {"rel": "publication_pdf", "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf"}], "name": "Kite: Building Conversational Bots from Mobile Apps", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. ", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Derives task graphs from demonstrated Android app interactions and layers a conversational question-answer interface over those graphs. The evaluation examines task-model extraction and developer use of the resulting bot templates.", "topics": ["mobile-navigation", "workflow", "human-in-the-loop"], "updated_at": null, "url": "https://doi.org/10.1145/3210240.3210339"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/inclusionAI/GUI-G2-7B/tree/main", "gated": false, "hub_id": "inclusionAI/GUI-G2-7B"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "cf53a868a010a20b165386edaaafed8918071c35a3babf1766fbb51f020cb4a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "c1f11fea30d3a64445e983a011db4efb9515a4221ad770e3a5223699e665eebe", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/inclusionAI/GUI-G2-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "cf53a868a010a20b165386edaaafed8918071c35a3babf1766fbb51f020cb4a1", "supports": ["licenses.weights"], "url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md"}], "id": "url:90eea6f6d099a3297850", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/inclusionAI/GUI-G2-7B"}], "name": "GUI-G2 7B", "notes": "The model card and four public safetensors shards establish this specific release. Apache-2.0 is declared for the model. The release targets element grounding; its grounding results do not establish long-horizon task-completion success.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md", "target_id": "arxiv:2507.15846", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL grounding checkpoint trained with Gaussian point and coverage rewards whose granularity adapts to interface-element scale.", "topics": ["grounding", "reinforcement-learning", "reward-models"], "updated_at": null, "url": "https://huggingface.co/inclusionAI/GUI-G2-7B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/xlangai/AgentNet/raw/main/README.md", "population": "AgentNet collection", "qualifier": "approximately", "unit": "tasks", "value": 22600}], "evidence_urls": ["https://huggingface.co/datasets/xlangai/AgentNet/raw/main/README.md"], "formats": ["JSONL", "image archives"], "modalities": ["screenshots", "instructions", "actions", "generated reasoning"], "notes": "The card distinguishes human task instructions from generated summaries, reasoning, reflections, and quality fields. Generated labels are not independent human quality acceptance records.", "splits": [], "supervision": ["computer-use trajectories", "synthesized reasoning"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/AgentNet/tree/main", "gated": false, "hub_id": "xlangai/AgentNet", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "f23c588ea9fb27210f9d605e164a32527a24298ca516df98272ff7ca7110eac6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/xlangai/AgentNet"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:40:46Z", "role": "primary", "sha256": "327a77856df259db0a0c2405478f75a0d0f2bf17c51806abc7c509ad82c3ad57", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/xlangai/AgentNet/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "f23c588ea9fb27210f9d605e164a32527a24298ca516df98272ff7ca7110eac6", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xlangai/AgentNet"}], "id": "url:964d70c40bfc5ca8e1ee", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/xlangai/AgentNet"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/AgentNet"}], "name": "AgentNet", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/xlangai/AgentNet", "target_id": "arxiv:2508.09123", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "22.6K human-annotated desktop computer-use tasks on Windows, macOS and Ubuntu, released with OpenCUA for training and evaluating vision-language-action models.", "topics": ["data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/xlangai/AgentNet"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "d86315819750e4a5fde12bce85a03c61cf68e704cb54146e6180027a37471b72", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.browserbase.com/"}], "id": "url:978d1dcd70fde8661b9e", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browserbase.com/"}], "name": "Browserbase Cloud", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://www.browserbase.com/", "target_id": "organization:browserbase", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Managed browser infrastructure for agent applications.", "topics": ["sandbox"], "updated_at": null, "url": "https://www.browserbase.com/"}
{"aliases": [], "authors": ["Sarah E. Chasins", "Maria Mueller", "Rastislav Bodik"], "details": {"publication": {"authors": ["Sarah E. Chasins", "Maria Mueller", "Rastislav Bodik"], "date_basis": "Supplied year precision: 2018. Rousillon entry: UIST 2018", "date_parts": [2018], "doi": "10.1145/3242587.3242661", "event_dates": null, "evidence_urls": ["https://schasins.com/papers/", "https://schasins.com/assets/papers/rousillon.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "This is browser interaction learned from demonstrations, rather than a generic server-side crawling library. Final proceedings page range and volume are not established. Manuscript pages 1-13 are not substituted for publisher pagination. Exact publication month/day is not inferred from conference dates.", "pages": null, "pdf_url": "https://schasins.com/assets/papers/rousillon.pdf", "publisher": null, "series": null, "title": "Rousillon: Scraping Distributed Hierarchical Web Data", "type": "conference-paper", "url": "https://doi.org/10.1145/3242587.3242661", "venue": "Proceedings of the 31st Annual ACM Symposium on User Interface Software and Technology", "volume": null}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:38Z", "role": "primary", "sha256": "6b7e584bb3862e0de51ca164d0f311c2a853e97bd63d56658ce62f0b49d86557", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://schasins.com/papers/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:12Z", "role": "primary", "sha256": "764dd1f5fe1db398188f403b4b1f6e3658e3238e4d8aac9ceed360dcad87e818", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "details.publication.url"], "url": "https://schasins.com/assets/papers/rousillon.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:12Z", "role": "primary", "sha256": "764dd1f5fe1db398188f403b4b1f6e3658e3238e4d8aac9ceed360dcad87e818", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "details.publication.url", "environments", "name", "scope", "summary", "topics"], "url": "https://schasins.com/assets/papers/rousillon.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:38Z", "role": "primary", "sha256": "6b7e584bb3862e0de51ca164d0f311c2a853e97bd63d56658ce62f0b49d86557", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue", "kind"], "url": "https://schasins.com/papers/"}], "id": "url:97c0b1896525632434e2", "identifiers": {"doi": "10.1145/3242587.3242661"}, "kind": "paper", "last_verified_at": "2026-10-17T09:55:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/3242587.3242661"}, {"rel": "publication_pdf", "url": "https://schasins.com/assets/papers/rousillon.pdf"}], "name": "Rousillon: Scraping Distributed Hierarchical Web Data", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. This is browser interaction learned from demonstrations, rather than a generic server-side crawling library.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Rousillon, which generalizes a demonstration of collecting one hierarchical data row into a browser-automation program that traverses related pages. The paper develops relation-selection and generalization methods and evaluates users building such scripts.", "topics": ["web-navigation", "recording", "automation"], "updated_at": null, "url": "https://doi.org/10.1145/3242587.3242661"}
{"aliases": [], "authors": [], "details": {"availability": "Public ungated JSON artifact is listed by the Hub manifest.", "dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/api/datasets/BlitherBoom/AutoDroid-V2-Dataset", "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_3_training/README.md"], "formats": ["JSON in ShareGPT conversation format"], "modalities": ["text"], "notes": "The card has license metadata only. The project training README supplies formatting and purpose. No sample count or split inventory is inferred from the training configuration.", "splits": [], "supervision": ["supervised conversations for AutoDroid-V2 fine-tuning"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset/tree/main", "gated": false, "hub_id": "BlitherBoom/AutoDroid-V2-Dataset", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset/raw/main/README.md"], "reason": "Publisher dataset-card license declaration; does not establish permissions for model weights or third-party application content."}}, "release_revision": "f041810dd32a2f2999647b7d4f4b8438482c01b8"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:33Z", "role": "primary", "sha256": "1b300edfbf44167da71639a56f8ab668dabd33769575fb313ebef5220fc1bddb", "supports": ["name", "kind", "identity", "release_status", "details.hub_id", "details.gated", "details.disabled", "details.release_revision", "details.availability", "details.download_url", "details.dataset_profile.modalities"], "url": "https://huggingface.co/api/datasets/BlitherBoom/AutoDroid-V2-Dataset"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:33Z", "role": "primary", "sha256": "d8d7a46d41a1a37fe4f0a5f637bf55c649310185329127d8a2204632e480be17", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "8ba37983a2c2da53d96a52696c07e4bdbea4e987a1a07b557935d614c3be46d2", "supports": ["summary", "topics", "scope", "notes", "details.dataset_profile", "relations"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_3_training/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "74508100843f8db6db6bb151e77ac1cc604a6b7ebd16d3225266d73434465136", "supports": ["environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/README_minimal.md"}], "id": "url:9a3506ca6ef034adb424", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T09:15:33Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card license declaration; does not establish permissions for model weights or third-party application content.", "declared": "mit", "retrieved_at": "2026-10-17T09:15:33Z", "url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset"}], "name": "AutoDroid V2 training data", "notes": "The public Hub manifest lists AutoDroid-V2-Dataset.json; no training records were downloaded. The card contains only an MIT metadata declaration, so the description and intended training use come from the official repository training guide. Sample count, split sizes and third-party-content rights remain unverified; the training command max_samples value is not treated as a dataset count.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_3_training/README.md", "target_id": "github:mobilellm/autodroid-v2", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "Conversation-formatted supervision released for fine-tuning AutoDroid V2, with the project documenting a ShareGPT-style JSON configuration.", "topics": ["data-generation", "imitation-learning", "mobile-navigation"], "updated_at": null, "url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/tree/main", "gated": false, "hub_id": "Yuqi-Zhou/GUI-G1-3B-v1"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "b7a4628a4d65ae226b50086075f3977b8f69ad12d15238a194675875bb9113e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "2f4b2084ca9cfd67f0952ea8c578c87f3a5d721064b5de3a191d7489373125aa", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Yuqi-Zhou/GUI-G1-3B-v1"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "b7a4628a4d65ae226b50086075f3977b8f69ad12d15238a194675875bb9113e8", "supports": ["licenses.weights"], "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md"}], "id": "url:9dc9d8d3b9a4b32d6f4b", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:30Z", "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1"}], "name": "GUI-G1 3B v1", "notes": "Two public safetensors shards and model/tokenizer configuration are present. The card declares MIT. Its inference example loads the v1 model but names a different 0.1K processor checkpoint; that example was not executed, and compatibility is not asserted.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md", "target_id": "arxiv:2505.15810", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-3B checkpoint released for R1-style visual grounding experiments, predicting interface target locations from screenshots and instructions.", "topics": ["grounding", "reinforcement-learning"], "updated_at": null, "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/openbmb/CAGUI/tree/main", "gated": false, "hub_id": "openbmb/CAGUI", "languages": ["zh"], "modalities": [], "size_category_reported": ["1K<n<10K"]}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "98699e13f7c7f5af812e1cbbd5e71eaceefa8aa73ca37167a772774a40818efd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/openbmb/CAGUI"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "98699e13f7c7f5af812e1cbbd5e71eaceefa8aa73ca37167a772774a40818efd", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/openbmb/CAGUI"}], "id": "url:a06622da7688e3718a7b", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "cc-by-nc-4.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/openbmb/CAGUI"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/openbmb/CAGUI"}], "name": "CAGUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/openbmb/CAGUI", "target_id": "arxiv:2506.01391", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Chinese Android GUI benchmark released with AgentCPM-GUI, with a grounding track for widgets and OCR text and an agent track for multi-step tasks in Chinese apps.", "topics": ["data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/openbmb/CAGUI"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/apurvaga/go-browse-wa/raw/main/README.md", "population": "train split", "qualifier": "exact", "unit": "step records", "value": 164533}], "evidence_urls": ["https://huggingface.co/datasets/apurvaga/go-browse-wa/raw/main/README.md", "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md"], "formats": ["Parquet"], "modalities": ["text", "accessibility-tree observations", "actions"], "notes": "The card count is flattened step records, not trajectories or independent tasks. Both successful and unsuccessful trajectories are included according to the project README; no license is declared.", "splits": ["train"], "supervision": ["action completion", "trajectory outcome"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/apurvaga/go-browse-wa/tree/main", "gated": false, "hub_id": "apurvaga/go-browse-wa", "languages": [], "modalities": ["tabular", "text"], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "494f4ac9ab9e94831dcf768be001b7284e30e00704674358baf3bb2060344361", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/apurvaga/go-browse-wa"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "18ab64833acf31113a83b2463359f1c7e37c4aaabfe11c33db2ba3e7d32818c6", "supports": ["details"], "url": "https://huggingface.co/datasets/apurvaga/go-browse-wa/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "0d4b28c993ccb977c7910ff3fecb5de29db364a7a7ad55523c8e052c9bae2566", "supports": ["summary", "environments", "relations"], "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "494f4ac9ab9e94831dcf768be001b7284e30e00704674358baf3bb2060344361", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/apurvaga/go-browse-wa"}], "id": "url:a13a187bc058fbb4f6af", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/apurvaga/go-browse-wa"}], "name": "Go-Browse-WA processed data", "notes": "The public release contains nine Parquet shards. The short card supplies schema and split metadata but no license declaration; the official README describes successful and unsuccessful trajectories processed for fine-tuning. The raw screenshot/HTML corpus is a separate release and is not claimed to be included here.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md", "target_id": "arxiv:2506.03533", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Flattened web-interaction training records with accessibility-tree observations, assistant completions, trajectory rewards and step-link metadata.", "topics": ["web-navigation", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/apurvaga/go-browse-wa"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/tree/main", "gated": false, "hub_id": "Hcompany/Holo3-35B-A3B"}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "0681f87b059110a4e5dff8955e9feca6c5ea118b326f729f9c0ed0fe49e42d03", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "bc5324dd6ceea9a0cc6bf0639b99d24cdd172af3c5d247bf923fdce1dc11f582", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo3-35B-A3B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "0681f87b059110a4e5dff8955e9feca6c5ea118b326f729f9c0ed0fe49e42d03", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/raw/main/README.md"}], "id": "url:a349aace3cfde25628a8", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B"}], "name": "Holo3 35B-A3B", "notes": "Fourteen public safetensors shards are present. The card declares Apache-2.0 for this release and describes a 35B-total/3B-active MoE architecture. No unpublished enterprise benchmark or training corpus is added solely from its performance claims.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A sparse Qwen3.5-based GUI-action checkpoint released for navigation and computer-use workflows.", "topics": ["grounding", "planning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B"}
{"aliases": [], "authors": ["Tom Yeh", "Tsung-Hsiang Chang", "Robert C. Miller"], "details": {"publication": {"authors": ["Tom Yeh", "Tsung-Hsiang Chang", "Robert C. Miller"], "date_basis": "Supplied month precision: 2009-10. Date Issued: October 2009", "date_parts": [2009, 10], "doi": "10.1145/1622176.1622213", "event_dates": null, "evidence_urls": ["https://dspace.mit.edu/entities/publication/604a3f79-451d-4186-a94e-d12690ffe40a"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Institutional abstract and published-version citation supply the retained evidence; the institution identifies an author final manuscript record. Exact publication day is not supplied. Operating-system-specific experiments are not asserted.", "pages": "183-192", "pdf_url": null, "publisher": null, "series": null, "title": "Sikuli: Using GUI screenshots for search and automation", "type": "conference-paper", "url": "https://doi.org/10.1145/1622176.1622213", "venue": "Proceedings of the 22nd Annual ACM Symposium on User Interface Software and Technology (UIST '09)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbaaf21982522b34857be014b39d6c43fedcda20f808efe40d04958c77243a5a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://dspace.mit.edu/entities/publication/604a3f79-451d-4186-a94e-d12690ffe40a"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbaaf21982522b34857be014b39d6c43fedcda20f808efe40d04958c77243a5a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://dspace.mit.edu/entities/publication/604a3f79-451d-4186-a94e-d12690ffe40a"}], "id": "url:a4cdb2ee0d07d3980651", "identifiers": {"doi": "10.1145/1622176.1622213"}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/1622176.1622213"}], "name": "Sikuli: Using GUI screenshots for search and automation", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. MIT identifies the retained content as an author final manuscript record. Its PDF download returned HTTP 429; no retry was made. The complete institutional abstract and bibliographic record were read.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces screenshot-based GUI search and visual scripting, using image patterns to locate interface targets and direct mouse and keyboard events. It evaluates screenshot queries and demonstrates automation and interactive-help tasks.", "topics": ["grounding", "automation", "screen-understanding"], "updated_at": null, "url": "https://doi.org/10.1145/1622176.1622213"}
{"aliases": ["url:4338ec0f1a67bdb6a139", "microsoft/Fara-7B", "https://huggingface.co/microsoft/Fara-7B", "url:d416331cdf9730afd6b9", "microsoft/fara-7b", "https://huggingface.co/microsoft/fara-7b"], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/microsoft/Fara-7b/tree/main", "gated": false, "hub_id": "microsoft/Fara-7B", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "92bfa3c1f34e7e76a35f72389a3b3dd9fdab5fcff0569c9fc8aabe331c703f07", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/microsoft/Fara-7b"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "92bfa3c1f34e7e76a35f72389a3b3dd9fdab5fcff0569c9fc8aabe331c703f07", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/Fara-7b"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:59Z", "role": "primary", "sha256": "cd1f83cd2ebf30d45501d1a31ecbe3bcb2039928d580c750733d2edb49916398", "supports": ["aliases"], "url": "https://huggingface.co/api/models/microsoft/Fara-7B"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:00Z", "role": "primary", "sha256": "cd1f83cd2ebf30d45501d1a31ecbe3bcb2039928d580c750733d2edb49916398", "supports": ["aliases"], "url": "https://huggingface.co/api/models/microsoft/Fara-7b"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:01Z", "role": "primary", "sha256": "cd1f83cd2ebf30d45501d1a31ecbe3bcb2039928d580c750733d2edb49916398", "supports": ["aliases"], "url": "https://huggingface.co/api/models/microsoft/fara-7b"}], "id": "url:a4da77ee63f879cc8e63", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/microsoft/Fara-7b"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/Fara-7b"}], "name": "Fara 7B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/microsoft/Fara-7b", "target_id": "github:microsoft/fara", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "Microsoft's Fara checkpoint, a Qwen2.5-VL-architecture image-text-to-text model with about 8.3B BF16 parameters, English, MIT-licensed and linked to arXiv paper 2511.19663.", "topics": ["planning", "web-navigation"], "updated_at": null, "url": "https://huggingface.co/microsoft/Fara-7b"}
{"aliases": ["url:2f67f5ad7954632dfc4e", "THUDM/webrl-llama-3.1-8b", "https://huggingface.co/THUDM/webrl-llama-3.1-8b"], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b/tree/main", "gated": false, "hub_id": "zai-org/webrl-llama-3.1-8b"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "4712446382f106a115a317d0ec6273a975545386b9cedf56946a96a56479dc9f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "6523b7f198d8debec2f3074190812955c6251590b61f8a6ea1bbb681b8ea128f", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/webrl-llama-3.1-8b"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "6523b7f198d8debec2f3074190812955c6251590b61f8a6ea1bbb681b8ea128f", "supports": ["identity"], "url": "https://huggingface.co/api/models/THUDM/webrl-llama-3.1-8b"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "4712446382f106a115a317d0ec6273a975545386b9cedf56946a96a56479dc9f", "supports": ["licenses.weights"], "url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:06Z", "role": "primary", "sha256": "96621f060dddec91453d8c2b451b3793a34b34810d82c7cd13ce4cb05ca90a76", "supports": ["aliases"], "url": "https://huggingface.co/api/models/THUDM/webrl-llama-3.1-8b"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:06Z", "role": "primary", "sha256": "96621f060dddec91453d8c2b451b3793a34b34810d82c7cd13ce4cb05ca90a76", "supports": ["aliases"], "url": "https://huggingface.co/api/models/zai-org/webrl-llama-3.1-8b"}], "id": "url:a6ee20fea67438d91e06", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:05:40Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T02:05:40Z", "url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b"}], "name": "WebRL Llama 3.1 8B", "notes": "The official Hub API resolves the former THUDM namespace to zai-org. Seven safetensors shards and configuration/tokenizer files are public and ungated. The card marks the weight license as other without naming complete terms, so no permissive weight license is inferred. The card’s evaluation prose also names GLM-4 despite the Llama checkpoint identity; no score is imported.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b/raw/main/README.md", "target_id": "arxiv:2411.02337", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A WebRL browser-policy checkpoint for performing actions in the five website environments of WebArena.", "topics": ["web-navigation", "reinforcement-learning"], "updated_at": null, "url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset/tree/main", "gated": false, "hub_id": "osunlp/GUI-Drag-dataset", "languages": [], "modalities": ["image"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "cdb8be8c4692a8ddcf8a66c1f0be666a9214718445925fb3c88615a0d641bf3a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/osunlp/GUI-Drag-dataset"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:54Z", "role": "primary", "sha256": "907c8a5169c180491ef8c26c9428b2c67bc4e26f472d6423ce3715cbe47f76ef", "supports": ["notes"], "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "cdb8be8c4692a8ddcf8a66c1f0be666a9214718445925fb3c88615a0d641bf3a", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/osunlp/GUI-Drag-dataset"}], "id": "url:aa28e1329a52e341d285", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset"}], "name": "GUI-Drag", "notes": "The card links its paper (arXiv 2601.06031) and code repository (OSU-NLP-Group/GUI-Drag). Neither is a reviewed catalog record yet, so no relation is recorded.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Data for learning text-dragging interactions in graphical interfaces.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset"}
{"aliases": [], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:51:44.325799+00:00", "role": "primary", "sha256": "1f0315e8758fe0edd97adb2c5253955dcdbe6d7bd8be393fe14dcd093483705f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://www.deeplearning.ai/courses/building-ai-browser-agents"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:41:10.403317+00:00", "role": "primary", "sha256": "0915d862638a308da6200d9603d668ab4d11bcabfa713ff0ccb82c7b21a5a92f", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://learn.deeplearning.ai/courses/building-ai-browser-agents/lesson/wwewm/building-a-simple-web-agent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:42:56.029732+00:00", "role": "primary", "sha256": "20d17ca47d1d16826f8e2736d6639d0cf11e826ef28c788187e16adcac9d7561", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://learn.deeplearning.ai/courses/building-ai-browser-agents/lesson/lot5j/building-an-autonomous-web-agent"}], "id": "url:aa693ca3e8948623f31e", "identifiers": {}, "kind": "learning", "last_verified_at": "2026-10-20T12:51:44.325799+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.deeplearning.ai/courses/building-ai-browser-agents"}], "name": "Building AI Browser Agents", "notes": "Review covers the public syllabus and complete published transcripts of the simple-agent and autonomous-agent lessons, not the whole course or its video frames. The first uses Playwright and structured HTML extraction; the second uses a MultiOn session API and shows missed steps, continuations and clarifications. Hosted notebooks and helper files were not retrieved, and current service access was not tested. No account, enrollment, terms acceptance, installation or execution occurred. Course and example-code reuse terms remain unknown.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Browser-agent course with worked lessons on structured web extraction and multi-step browser interaction, followed by AgentQ search and self-correction concepts.", "topics": ["web-navigation", "planning"], "updated_at": null, "url": "https://www.deeplearning.ai/courses/building-ai-browser-agents"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/alibabagroup/CMGUI/tree/main", "gated": false, "hub_id": "alibabagroup/CMGUI", "languages": ["zh"], "modalities": [], "size_category_reported": ["100K<n<1M"]}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "aa1e063e73269acd0cb06adcef67d6230f0fe9498ed361d92234053dcf035685", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/alibabagroup/CMGUI"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:26Z", "role": "primary", "sha256": "6f43b9e5f2c783f91cc3271f51db6a9a521c29f21954a2a30ad0cfa186e1c1b7", "supports": ["relations"], "url": "https://huggingface.co/datasets/alibabagroup/CMGUI/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "aa1e063e73269acd0cb06adcef67d6230f0fe9498ed361d92234053dcf035685", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/alibabagroup/CMGUI"}], "id": "url:ab26f9ef56d651cee587", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/alibabagroup/CMGUI"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/alibabagroup/CMGUI"}], "name": "CMGUI", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/alibabagroup/CMGUI/raw/main/README.md", "target_id": "arxiv:2603.08533", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "18k trajectories with 98k steps from more than 50 Chinese mobile apps in domains such as e-commerce and social media, for building GUI agents on Chinese apps.", "topics": ["data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/alibabagroup/CMGUI"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["candidate-decision preference"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/raw/main/README.md", "population": "Hub test split", "qualifier": "exact", "unit": "rows", "value": 776}], "evaluation": "Compare reward-model preferences with the provided chosen/rejected decision labels; no independent score was reproduced.", "evidence_urls": ["https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/raw/main/README.md", "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"], "focus": ["web process-reward evaluation"], "notes": "The current Hub row count and project case-count claim use unresolved counting units. Candidate lists mean a row must not be assumed to equal one task or one pair.", "observations": ["page screenshots", "text observations", "action history", "task checklists"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/tree/main", "gated": false, "hub_id": "LangAGI-Lab/WebRewardBench"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "0968f60eea10b95f79cd16ad3746b0e0b1acaeb0e97c45856f709d53a14dd73a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "db06c66b8cf5687008b0ad0d9ae269fcc27936887d238cfe2bdeb52151e1683e", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/LangAGI-Lab/WebRewardBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "07431058d8b9a71b80e75b7a74a40a8346c1503eec240628459cc4d4e2f0802f", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"}], "id": "url:ac6d0f80765f5a32255a", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench"}], "name": "WebRewardBench", "notes": "Two public Parquet shards are present. Card metadata reports 776 test rows, whereas the project describes 1,000-plus test cases without reconciling that count to rows. Both observations are retained as a grain/coverage caveat rather than treating either as a confirmed number of independent tasks. No data license is declared.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/raw/main/README.md", "target_id": "arxiv:2505.15277", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "An offline preference-evaluation set for web process reward models, pairing page and trajectory context with preferred and rejected candidate decisions.", "topics": ["web-navigation", "reward-models", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/xlangai/OpenCUA-32B/tree/main", "gated": false, "hub_id": "xlangai/OpenCUA-32B"}, "environments": ["desktop", "linux", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "151217eb96e4ab22a47c98766052e5cfd76497a2a9197cc125c0237f3fbc59f6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "a3be2e1a6cd39ee89809baa9504813c4a6a9a03bbd916e82924cc0f89cb5da33", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-32B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "151217eb96e4ab22a47c98766052e5cfd76497a2a9197cc125c0237f3fbc59f6", "supports": ["licenses.weights"], "url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md"}], "id": "url:ad7f176848fd84de64e2", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/xlangai/OpenCUA-32B"}], "name": "OpenCUA 32B", "notes": "The public manifest contains 64 safetensors shards. The card declares MIT for this model and names its Qwen2.5-VL base; upstream terms remain separate. This is the 32B release, distinct from the existing 72B preview. Family-level or 72B benchmark scores are not transferred to it.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md", "target_id": "arxiv:2508.09123", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-based computer-use checkpoint that combines screenshot history with planning and grounded executable actions.", "topics": ["grounding", "planning", "self-reflection"], "updated_at": null, "url": "https://huggingface.co/xlangai/OpenCUA-32B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/mllmTeam/DroidCall/tree/main", "gated": false, "hub_id": "mllmTeam/DroidCall", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md"], "reason": "The Hub dataset card declares Apache-2.0 in its metadata header (license: apache-2.0). The companion repository's MIT code license is a separate declaration."}}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "08a6c5cfa655e1ecd774a41a9a815e845ad0afd47b289378bc1d6ee66c81dda9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "c7804987d56b03314a4c48e3ddf2e5819fc0d9c83cc65ba54a981a2f4a584dd9", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/mllmTeam/DroidCall"}], "id": "url:ae8f5208eea9e7f1f606", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/mllmTeam/DroidCall"}], "name": "DroidCall data", "notes": "The release provides JSONL data and API annotations. This is Android intent-calling supervision rather than screenshot-and-click trajectories, so it is marked adjacent. The dataset card declares Apache-2.0; the companion repository’s MIT code license is a separate declaration.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md", "target_id": "arxiv:2412.00402", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "Instruction-to-function-call examples for learning Android intent invocation, with training, testing and API-annotation files for mobile task execution.", "topics": ["mobile-navigation", "tool-integration", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/mllmTeam/DroidCall"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/mPLUG/GUI-Owl-7B/tree/main", "gated": false, "hub_id": "mPLUG/GUI-Owl-7B"}, "environments": ["android", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "a0e6dd76aee6f138c15f244f4e2e8ba650b8829295397fdbfa839a8edf852b57", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "4822eab7fd35e583d52db7369d7d929d69dbde66b27c9da3d370cad9ec06ec89", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/mPLUG/GUI-Owl-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "a0e6dd76aee6f138c15f244f4e2e8ba650b8829295397fdbfa839a8edf852b57", "supports": ["licenses.weights"], "url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md"}], "id": "url:aeb32973694f91ad9920", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/mPLUG/GUI-Owl-7B"}], "name": "GUI-Owl 7B", "notes": "Five public safetensors shards are present. MIT is declared in this model card. This record identifies the original 7B release; other sizes and later GUI-Owl generations are not aliases or assumed to share its exact terms. The online demo is still marked forthcoming.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md", "target_id": "arxiv:2508.15144", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-based GUI checkpoint released with Mobile-Agent-v3 for grounded actions and specialized roles in its agent framework.", "topics": ["grounding", "action-prediction", "planning"], "updated_at": null, "url": "https://huggingface.co/mPLUG/GUI-Owl-7B"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "fb3b4dbc435158c256abae36cf092172284c4d6fd53fa556499cb7b0730a5bc6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://openai.com/index/introducing-chatgpt-agent/"}], "id": "url:b074ccdb2d9ec0277b65", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/introducing-chatgpt-agent/"}], "name": "ChatGPT Agent", "notes": "The launch page identifies itself as outdated. This entry records the product introduction, not current availability, limits, or pricing.", "published_at": null, "relations": [{"evidence_url": "https://openai.com/index/introducing-chatgpt-agent/", "target_id": "organization:openai", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Hosted agent combining browser interaction, tools, and computer-based tasks.", "topics": ["workflow"], "updated_at": null, "url": "https://openai.com/index/introducing-chatgpt-agent/"}
{"aliases": [], "authors": [], "details": {"availability": "Ungated v1.0 JSON task-definition file.", "benchmark_protocol": {"actions": ["mobile GUI actions through a separately configured agent"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md", "population": "MobBench tasks used in the paper", "qualifier": "exact", "unit": "tasks", "value": 50}], "evaluation": "The publisher reports a fifty-task real-device evaluation; the inspected release does not establish an automatic correctness judge.", "evidence_urls": ["https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md"], "focus": ["mobile task completion"], "notes": "Task definitions are distinct from device images, agent execution and automated scoring.", "observations": ["Android device interface"], "setting": "interactive"}, "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md"], "reason": "Dataset-card declaration for the MobBench task-definition release only."}}, "release_revision": "28941d10c49a16efb991d4157e512e14204bf3d6"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "231a8f483256c774840c99383ab6536520a0aed0f04d62b4d64f1f59477b906e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol", "licenses.data"], "url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "4e7225f956d7cee98fe671e7bfd91fc407d87dd805d8f0c11c3b0c839714ccf3", "supports": ["identity", "release_status", "details.release_revision"], "url": "https://huggingface.co/api/datasets/OpenDFM/MobA-MobBench"}], "id": "url:b53e03a6e4ca4226d126", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T12:32:07Z", "licenses": {"code": null, "data": {"basis": "Dataset-card declaration for the MobBench task-definition release only.", "declared": "cc-by-nc-sa-4.0", "retrieved_at": "2026-10-17T12:32:07Z", "url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench"}], "name": "MobBench task definitions", "notes": "The ungated manifest lists MobBench_tasks_release_v1.0.json and bilingual cards. The card reports fifty tasks; no sample records were downloaded to recount them. This release supplies task definitions and agent setup guidance, not an independently verified automatic judge or a bundled Android environment. The data card declares CC-BY-NC-SA-4.0; code and model licenses are separate.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md", "target_id": "arxiv:2410.13757", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A public set of real-device mobile task definitions used to evaluate MobA, with Chinese and English task material packaged in a versioned JSON release.", "topics": ["evaluation", "mobile-navigation"], "updated_at": null, "url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench"}
{"aliases": [], "authors": ["Biplab Deka", "Zifeng Huang", "Chad Franzen", "Joshua Hibschman", "Daniel Afergan", "Yang Li", "Jeffrey Nichols", "Ranjitha Kumar"], "details": {"publication": {"authors": ["Biplab Deka", "Zifeng Huang", "Chad Franzen", "Joshua Hibschman", "Daniel Afergan", "Yang Li", "Jeffrey Nichols", "Ranjitha Kumar"], "date_basis": "Supplied day precision: 2017-10-20. PDF cover, Published: 20 October 2017 The 18 January 2026 download date and October 22-25 conference dates are not publication dates.", "date_parts": [2017, 10, 20], "doi": "10.1145/3126594.3126651", "event_dates": null, "evidence_urls": ["https://danafergan.com/publications/deka2017rico.pdf", "https://dev.interactionmining.org/archive/rico"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The paper is a distinct resource from the already cataloged Rico dataset (url:0b134b082a7bd507fc6a). The publisher PDF describes 9.7k apps and 72k screens; the project page describes 9.3k apps and 66k screens. The annotation intentionally supplies no merged size claim. No volume number is supplied for this proceedings article.", "pages": "845-854", "pdf_url": "https://danafergan.com/publications/deka2017rico.pdf", "publisher": null, "series": null, "title": "Rico: A Mobile App Dataset for Building Data-Driven Design Applications", "type": "conference-paper", "url": "https://doi.org/10.1145/3126594.3126651", "venue": "UIST '17: Proceedings of the 30th Annual ACM Symposium on User Interface Software and Technology", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T10:01:32Z", "role": "primary", "sha256": "a1ec58c526b152c052c5c72f9a62725f1026b5f0e4f45c8fe85fce284295da84", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://danafergan.com/publications/deka2017rico.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:23Z", "role": "primary", "sha256": "fb00401fe531ca3aed078dccc57147bcaad33207dacf7cd4cfd4af14f1487ff5", "supports": ["details.publication.notes"], "url": "https://dev.interactionmining.org/archive/rico"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T10:01:32Z", "role": "primary", "sha256": "a1ec58c526b152c052c5c72f9a62725f1026b5f0e4f45c8fe85fce284295da84", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://danafergan.com/publications/deka2017rico.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:23Z", "role": "primary", "sha256": "fb00401fe531ca3aed078dccc57147bcaad33207dacf7cd4cfd4af14f1487ff5", "supports": ["details.publication.notes"], "url": "https://dev.interactionmining.org/archive/rico"}], "id": "url:b560b5d6abaab98483f6", "identifiers": {"doi": "10.1145/3126594.3126651"}, "kind": "paper", "last_verified_at": "2026-10-17T10:01:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/3126594.3126651"}, {"rel": "publication_pdf", "url": "https://danafergan.com/publications/deka2017rico.pdf"}], "name": "Rico: A Mobile App Dataset for Building Data-Driven Design Applications", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The paper is a distinct resource from the already cataloged Rico dataset (url:0b134b082a7bd507fc6a). The publisher PDF describes 9.7k apps and 72k screens; the project page describes 9.3k apps and 66k screens. The annotation intentionally supplies no merged size claim.", "published_at": "2017-10-20", "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Rico, collecting Android interface designs and interaction traces through crowdsourced use followed by automated exploration. The paper demonstrates UI-layout retrieval using learned representations and describes visual, textual, structural, and interaction data.", "topics": ["data-generation", "screen-understanding", "ui-design"], "updated_at": null, "url": "https://doi.org/10.1145/3126594.3126651"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md", "https://huggingface.co/api/datasets/xlangai/aguvis-stage2", "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md", "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage2.yaml"], "formats": ["JSON annotations", "ZIP image packages", "multipart TAR image packages"], "modalities": ["GUI screenshots", "natural-language instructions", "action trajectories"], "notes": "Stage 2 trajectory release, Hub revision bd6f1ed4, is distinct from Stage 1 grounding. The public manifest lists source-specific JSON levels and image ZIP/multipart TAR packages. The training recipe uses all rows for some inputs and random:33% for others; several l2 paths occur twice while the manifest also contains l3 files. Recipe rows/levels are neither independent datasets nor disjoint train/test splits. No example/action/trajectory count is inferred. Apache-2.0 is the card declaration, with separate underlying source terms.", "splits": ["training"], "supervision": ["GUI trajectory supervised fine-tuning", "planning and reasoning with inner monologue"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/aguvis-stage2/tree/main", "gated": false, "hub_id": "xlangai/aguvis-stage2"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "550293f00905615e19e63ac896818f9875dad38fea0ffe196c6302a29719aa72", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "details.dataset_profile"], "url": "https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "f030264a7a8c7485dc7cf15ca554d84450cc492d2ac90e5965810a0207ef370b", "supports": ["identity", "release_status", "details", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/xlangai/aguvis-stage2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dd9f36b818c1b7b6bd772ba8575f6c0ebab6d9e0d307edcc2f196f3e4ba502e", "supports": ["environments", "summary", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:11Z", "role": "primary", "sha256": "8c1b39e9d16067018ce845fd4e996f34e2b47346c342b2ba26b030f4a5338b66", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage2.yaml"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "550293f00905615e19e63ac896818f9875dad38fea0ffe196c6302a29719aa72", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md"}], "id": "url:b5e2127e07f692d516b9", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/aguvis-stage2"}], "name": "Aguvis stage 2 trajectory data", "notes": "The manifest contains trajectory annotations and archives associated with mobile, desktop and web sources. The card declares Apache-2.0; underlying source content may retain separate terms. This stage has a different training purpose and files from the grounding release. No archive or training sample was downloaded.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md", "target_id": "arxiv:2412.04454", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The second-stage Aguvis collection of interface trajectories for learning action sequences, planning and reasoning after grounding pretraining.", "topics": ["planning", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/xlangai/aguvis-stage2"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xwk123/Mobile3M/raw/3df26694ff9186604eb44a8e633894d3ece67ed9/README.md", "https://huggingface.co/api/datasets/xwk123/Mobile3M", "https://raw.githubusercontent.com/XiaoMi/mobilevlm/HEAD/README.md"], "formats": ["7z app-graph archives", "multipart 7z files"], "modalities": [], "notes": "Official README calls this a partial training release and says app data will be uploaded gradually. Its training-JSON and test-data links point to separate Hub repositories; those artifacts are not counted as files or splits of xwk123/Mobile3M. Public revision 3df26694 lists app-graph archives; their names do not establish a three-million-example downloadable inventory. Precise inner record/modalities schema is not inspected. The card YAML says Apache-2.0 while its prose and official README assign CC-BY-NC-SA-4.0 to data and Apache to code; the existing data-license value remains unknown with the conflict documented.", "splits": ["partial training release"], "supervision": ["mobile UI pretraining for MobileVLM"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xwk123/Mobile3M/tree/main", "gated": false, "hub_id": "xwk123/Mobile3M", "languages": [], "license_review": {"data": {"evidence_urls": ["https://huggingface.co/api/datasets/xwk123/Mobile3M"], "reason": "The retained first-party API has Apache-2.0 card metadata but its publisher description explicitly assigns Apache to source code and CC-BY-NC-SA-4.0 to the dataset. A single permissive dataset declaration is unresolved."}}, "modalities": ["image"], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "356a8dca28fa3dc46bc05b13eaf04a3db273e63c06be4fd15d0e87e0343c0824", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/xwk123/Mobile3M"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:10:01Z", "role": "primary", "sha256": "8a097dbc7384d3743f262dc0164c176ed577bf720cbbc92ff8b3cd820d1aed4f", "supports": ["details.dataset_profile", "relations"], "url": "https://huggingface.co/datasets/xwk123/Mobile3M/raw/3df26694ff9186604eb44a8e633894d3ece67ed9/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "e8a92a783b6b65cacedf105e5ede583668f8774f2a81a2005dc6897da6fa88ef", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/XiaoMi/mobilevlm/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "356a8dca28fa3dc46bc05b13eaf04a3db273e63c06be4fd15d0e87e0343c0824", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xwk123/Mobile3M"}], "id": "url:b80bfa0a4939a493a11f", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xwk123/Mobile3M"}], "name": "Mobile3M", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/xwk123/Mobile3M/raw/3df26694ff9186604eb44a8e633894d3ece67ed9/README.md", "target_id": "arxiv:2409.14818", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Pre-training corpus for Xiaomi's MobileVLM, distributed as per-app graph archives for mobile apps such as Baidu Map, QQ Music, Ctrip and Vipshop.", "topics": ["screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/xwk123/Mobile3M"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/Hcompany/DragOn/tree/main", "gated": false, "hub_id": "Hcompany/DragOn", "languages": ["en", "fr"], "modalities": [], "size_category_reported": ["1M<n<10M"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "08b5529ce01781143147d56e2278aec2e5bdd972ba7e2fae6ba7e65e62688d40", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/Hcompany/DragOn"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:54Z", "role": "primary", "sha256": "e7ef94cafcfd155961beec373f2f4e727715b7d517a819f52b5407fc2dec4009", "supports": ["notes"], "url": "https://huggingface.co/datasets/Hcompany/DragOn/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "08b5529ce01781143147d56e2278aec2e5bdd972ba7e2fae6ba7e65e62688d40", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Hcompany/DragOn"}], "id": "url:b854ff53e497da3a5ad8", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/Hcompany/DragOn"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Hcompany/DragOn"}], "name": "DragOn", "notes": "The card cites a workshop paper titled \"DragOn: A Drag-Grounding Benchmark and Training Dataset for GUI Agents\" without an arXiv link. arXiv 2606.06322 has a different title, so no paper relation is recorded.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Drag-grounding training and evaluation data pairing each screenshot instruction with start and end boxes, across text highlighting, spreadsheet cells, slide resizing and sliders.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/datasets/Hcompany/DragOn"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "5eacfa15c116ddeb02f6fd731fbb2e9856a933d743870fe41d947ef3b01aeb6f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://steel.dev/"}], "id": "url:b940a884f200a916a61b", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://steel.dev/"}], "name": "Steel Cloud", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://steel.dev/", "target_id": "organization:steel", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Managed browser infrastructure associated with Steel Browser.", "topics": ["sandbox"], "updated_at": null, "url": "https://steel.dev/"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/KDEGroup/UI-AGILE/tree/main", "gated": false, "hub_id": "KDEGroup/UI-AGILE"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "75aed1ebff7dd54ab1dc0112fe524a118d9e8845c2e079080d2bb7954f7ab36e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "a52032bb185885d47466f0bea26c15ccefeb8d6ff89f3bacc5eb4192c35d45f7", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/KDEGroup/UI-AGILE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:43:19Z", "role": "primary", "sha256": "ce35413bc1ef28338bc6fd9f132fa271a4f5e0bf23cecd7346f54adf0a964fd7", "supports": ["environments"], "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "75aed1ebff7dd54ab1dc0112fe524a118d9e8845c2e079080d2bb7954f7ab36e", "supports": ["licenses.weights"], "url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md"}], "id": "url:b94e285c3f757e2d0f3a", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:30Z", "url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/KDEGroup/UI-AGILE"}], "name": "UI-AGILE checkpoint", "notes": "Four safetensors shards and model/tokenizer files are public. The card declares MIT for these weights; the companion dataset has no declared license and is separate. The short card does not state a parameter count, so this entry does not infer one from shard sizes.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md", "target_id": "arxiv:2507.22025", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The released UI-AGILE model for GUI grounding with reward shaping, crop-based resampling and decomposed selection during inference.", "topics": ["grounding", "reinforcement-learning", "efficiency"], "updated_at": null, "url": "https://huggingface.co/KDEGroup/UI-AGILE"}
{"aliases": [], "authors": ["Toby Jia-Jun Li", "Igor Labutov", "Xiaohan Nancy Li", "Xiaoyi Zhang", "Wenze Shi", "Wanling Ding", "Tom M. Mitchell", "Brad A. Myers"], "details": {"publication": {"authors": ["Toby Jia-Jun Li", "Igor Labutov", "Xiaohan Nancy Li", "Xiaoyi Zhang", "Wenze Shi", "Wanling Ding", "Tom M. Mitchell", "Brad A. Myers"], "date_basis": "Supplied year precision: 2018. APPINITE publication citation: VL/HCC 2018", "date_parts": [2018], "doi": null, "event_dates": null, "evidence_urls": ["https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf", "https://www.cs.cmu.edu/~NatProg/sugilite.html", "https://toby.li/publications/"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The base SUGILITE framework and later PUMICE work are distinct publications; this paper focuses on resolving data-description ambiguity in demonstrations. Author-hosted manuscript and institutional publication lists supply the retained evidence. DOI, final page range, and exact publication day remain unknown.", "pages": null, "pdf_url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf", "publisher": null, "series": null, "title": "APPINITE: A Multi-Modal Interface for Specifying Data Descriptions in Programming by Demonstration Using Natural Language Instructions", "type": "conference-paper", "url": "https://ieeexplore.ieee.org/document/8506506", "venue": "Proceedings of the 2018 IEEE Symposium on Visual Languages and Human-Centric Computing", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:46Z", "role": "primary", "sha256": "6e05fd08b187adba32148e0835ef6b9e20cd6a3ba5e0de6e2778eb18c02788ee", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title"], "url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:22Z", "role": "primary", "sha256": "32ee05ff1a91ed4a22d4bc156a462872aa4e28f5c79abfbce5dbf3655f412033", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://www.cs.cmu.edu/~NatProg/sugilite.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:20Z", "role": "primary", "sha256": "adfcec73830b612b181d4846ace3fbaab1768e906715c1d90fe657e23fe4d985", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://toby.li/publications/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:46Z", "role": "primary", "sha256": "6e05fd08b187adba32148e0835ef6b9e20cd6a3ba5e0de6e2778eb18c02788ee", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "environments", "name", "scope", "summary", "topics"], "url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:22Z", "role": "primary", "sha256": "32ee05ff1a91ed4a22d4bc156a462872aa4e28f5c79abfbce5dbf3655f412033", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue", "kind"], "url": "https://www.cs.cmu.edu/~NatProg/sugilite.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:20Z", "role": "primary", "sha256": "adfcec73830b612b181d4846ace3fbaab1768e906715c1d90fe657e23fe4d985", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://toby.li/publications/"}], "id": "url:ba2c664c1b3c6f1b172f", "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:54:46Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://ieeexplore.ieee.org/document/8506506"}, {"rel": "publication_pdf", "url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf"}], "name": "APPINITE: A Multi-Modal Interface for Specifying Data Descriptions in Programming by Demonstration Using Natural Language Instructions", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The base SUGILITE framework and later PUMICE work are distinct publications; this paper focuses on resolving data-description ambiguity in demonstrations.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces APPINITE, a conversational interface for specifying the intended targets of demonstrated Android actions. It constructs semantic UI representations and uses an on-screen overlay to preview and refine executable data descriptions.", "topics": ["grounding", "human-in-the-loop", "mobile-navigation"], "updated_at": null, "url": "https://ieeexplore.ieee.org/document/8506506"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md", "https://huggingface.co/api/datasets/xTRam1/plan-and-act-data"], "formats": ["JSON", "JSONL"], "modalities": ["text", "plans", "actions"], "notes": "The release contains planner and actor files for two web environments. No sample count is inferred from the package-file manifest. Card loading examples have repository/path inconsistencies.", "splits": ["train (QwQ-32B files)", "validation (QwQ-32B files)"], "supervision": ["high-level plans", "action execution"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data/tree/main", "gated": false, "hub_id": "xTRam1/plan-and-act-data"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:29Z", "role": "primary", "sha256": "2ec8f4a03b8331bdd2ba8d6f6800d6dea9d955cfd9e7c1d2168ecb64deb25448", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "a98f98b7d4ceb81b00f217983ff271257f2f1e7d2c1a09836b872b2350d79ae9", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xTRam1/plan-and-act-data"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:02:29Z", "role": "primary", "sha256": "2ec8f4a03b8331bdd2ba8d6f6800d6dea9d955cfd9e7c1d2168ecb64deb25448", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md"}], "id": "url:ba6453cfdae0eb049ced", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:02:29Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:02:29Z", "url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data"}], "name": "Plan-and-Act training data", "notes": "The public manifest exposes JSON/JSONL data for planner and actor roles, including train and validation variants. The card declares MIT. Its loading examples use a different Hub repository identifier and omit the data/ prefix for one file; use the inspected xTRam1 manifest rather than copying those examples verbatim. Third-party site content retains its own terms.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md", "target_id": "arxiv:2503.09572", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Planner and action-executor training examples for WebArena-Lite and WebVoyager, with model-specific processed files and raw trajectory annotations.", "topics": ["planning", "web-navigation", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/GD-ML/Code2World/tree/3e48194a8228570164ecd510f9d83ac839e9e931", "gated": false, "hub_id": "GD-ML/Code2World", "hub_revision": "3e48194a8228570164ecd510f9d83ac839e9e931", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/GD-ML/Code2World/raw/3e48194a8228570164ecd510f9d83ac839e9e931/README.md"], "reason": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate."}}, "repository_url": "https://github.com/AMAP-ML/Code2World"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:15Z", "role": "primary", "sha256": "befc89b09d099b86bda891141bca32b4a13b650c64d9b74ca18a9431e76f6c06", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations", "licenses.weights"], "url": "https://huggingface.co/GD-ML/Code2World/raw/3e48194a8228570164ecd510f9d83ac839e9e931/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:22Z", "role": "primary", "sha256": "4c91d95f7533cb5550447c95bfb12d6064fd510ac01cba45ebb8d7fca305fdc9", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url"], "url": "https://huggingface.co/api/models/GD-ML/Code2World"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "26c14a14166d3b747454d1924ecdf6d96652ac857ae6ad488630dc843036a777", "supports": ["environments", "details.repository_url"], "url": "https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md"}], "id": "url:bc3deaf8eeecf569e73b", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-18T04:35:15Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:15Z", "url": "https://huggingface.co/GD-ML/Code2World/raw/3e48194a8228570164ecd510f9d83ac839e9e931/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/GD-ML/Code2World"}], "name": "Code2World-8B checkpoint", "notes": "Card identifies Code2World-8B and Qwen3-VL-8B-Instruct. Public files include four Safetensors shards, an index and rendering/prompt helper scripts. The GitHub adapter targets AndroidWorld; model-generated HTML rendering and model loading were not executed. MIT is recorded as the weight declaration, without transferring it to the separate GitHub repository’s code or all training sources.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/GD-ML/Code2World/raw/3e48194a8228570164ecd510f9d83ac839e9e931/README.md", "target_id": "arxiv:2602.09856", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Released GUI world-model checkpoint that turns a screenshot and proposed action into HTML for rendering a predicted next screen.", "topics": ["world-models", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/GD-ML/Code2World"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["click coordinates"], "counts": [], "evaluation": "The inspected Jedi evaluation adapter reads the v2 mobile/desktop/web annotation files, scores a point against the target box, and reports overall, platform, and text/icon accuracy. This is a named compatible evaluator implementation; the OS-Atlas dataset card itself supplies no evaluator specification.", "evidence_urls": ["https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2", "https://osatlas.github.io/", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_screenspot_v2_jedi.py", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"], "focus": ["single-step GUI grounding on the ScreenSpot-v2 release"], "notes": "The public Hub revision 5efbb1f1 contains three platform JSON annotation files and an image ZIP; its card only declares Apache-2.0. No sample total or equivalence to the original ScreenSpot test count is asserted. Keep v2 separate from v1 and retain evaluator/version and coordinate conversion when comparing results. Files were listed, not downloaded as data.", "observations": ["screenshot", "instruction"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2/tree/main", "gated": false, "hub_id": "OS-Copilot/ScreenSpot-v2", "languages": [], "modalities": ["image"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "7feddadc306e384f8df6a3d9c8afc0c776cb3734c4c7fe3eb842eab85537f408", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:05Z", "role": "primary", "sha256": "94547cb88756cf6a36bfb86d5fe9bdca1e40d7edb9e452c33cb462bbe760fed0", "supports": ["details.benchmark_protocol"], "url": "https://osatlas.github.io/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "0e8772bc8fd5a25c50d3c2f6b0caee05ed2ab4659d46d99d60adf2b5ab29e997", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_screenspot_v2_jedi.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "b85f5685ce6c200ea648471666177859693d7f40cf50c12d08765ae68580de4a", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "7feddadc306e384f8df6a3d9c8afc0c776cb3734c4c7fe3eb842eab85537f408", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2"}], "id": "url:bd7214de6d98917c2350", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"}], "name": "ScreenSpot v2", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Second version of the ScreenSpot grounding test set, published by OS-Copilot as separate desktop, mobile and web annotation files with a shared screenshot archive.", "topics": ["grounding", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"}
{"aliases": [], "authors": [], "details": {"availability": "Ungated JSON conversation shards and ranking-file manifest.", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "train", "qualifier": "exact", "unit": "conversations", "value": 600}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-task", "qualifier": "exact", "unit": "conversations", "value": 34}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-website", "qualifier": "exact", "unit": "conversations", "value": 42}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-subdomain", "qualifier": "exact", "unit": "conversations", "value": 44}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "train", "qualifier": "exact", "unit": "dialogue turns", "value": 2896}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-task", "qualifier": "exact", "unit": "dialogue turns", "value": 191}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-website", "qualifier": "exact", "unit": "dialogue turns", "value": 218}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-subdomain", "qualifier": "exact", "unit": "dialogue turns", "value": 216}], "evidence_urls": ["https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md"], "formats": ["JSON", "auxiliary PKL"], "modalities": ["dialogue text", "HTML", "element candidates", "actions"], "notes": "Conversation, dialogue-turn and action-step units differ. The released file manifest is inspected; underlying examples were not downloaded.", "splits": ["train", "test_task", "test_website", "test_subdomain"], "supervision": ["multi-turn web instruction following", "element/action prediction"]}, "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md"], "reason": "Publisher dataset-card declaration for the multi-turn annotation release."}}, "release_revision": "5ae9a3bedf431f54c2218fbb8ee474e5efbda3a8"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "8685e0a26a06b3339b5c078632a7b0ade1ead40e6b586ef1ff3676414699a4bc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "details.dataset_profile", "licenses.data"], "url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "4455e6daf1151694c03b2a1ebd4a7af40b5b079b0568706a1783a0222383bb1d", "supports": ["identity", "release_status", "details.release_revision"], "url": "https://huggingface.co/api/datasets/magicgh/MT-Mind2Web"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "id": "url:bdfb50330deb97893ce4", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T12:34:10Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for the multi-turn annotation release.", "declared": "cc-by-4.0", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web"}], "name": "MT-Mind2Web", "notes": "The ungated Hub manifest contains train, test-task, test-website and test-subdomain JSON shards plus a candidate-ranking pickle file. Counts distinguish conversation sessions from dialogue turns; neither is a screenshot or action-step count. No pickle or sample data was loaded. The card declares CC-BY-4.0 and identifies the original Mind2Web interactions as its source.", "published_at": null, "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2402.15057", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Multi-turn web-navigation conversations constructed from Mind2Web interactions, with dialogue turns, HTML snapshots, element candidates and click/type/select action sequences.", "topics": ["web-navigation", "imitation-learning", "memory"], "updated_at": null, "url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web"}
{"aliases": [], "authors": ["Fei Tang", "Zhangxuan Gu", "Zhengxi Lu", "Shangzhan Zhang", "Zhengwen Zeng", "Shuheng Shen", "Changhua Meng", "Yuchen Yan", "Wenqi Zhang", "Yongliang Shen", "Weiming Lu", "Yueting Zhuang"], "details": {"publication": {"authors": ["Fei Tang", "Zhangxuan Gu", "Zhengxi Lu", "Shangzhan Zhang", "Zhengwen Zeng", "Shuheng Shen", "Changhua Meng", "Yuchen Yan", "Wenqi Zhang", "Yongliang Shen", "Weiming Lu", "Yueting Zhuang"], "date_basis": "Explicit publisher metadata: month = {June}, year = {2026}; precision=month.", "date_parts": [2026, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html", "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf"], "identity_basis": "No title/DOI/canonical URL match in the current paper catalog or raw archive; closest substantive same-team works were reread and have different contributions.", "issue": null, "notes": "CVF open-access accepted version; PDF states that the final proceedings version is on IEEE Xplore.", "pages": "13007-13016", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf", "publisher": null, "series": null, "title": "GUI-SAGE: Enhancing GUI Automation with Self-Explanatory Learning", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "1d5ed4e24cf9763cc43f0a8ce14f0e98c01504c73578e5632cea573dfb7da647", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:56:04+00:00", "role": "primary", "sha256": "244f40d77a40dfd184e78cf8e1e2dd9e469d3e594a8ce214279c2ce17b8f1012", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "1d5ed4e24cf9763cc43f0a8ce14f0e98c01504c73578e5632cea573dfb7da647", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "scope", "summary", "topics"], "url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:56:04+00:00", "role": "primary", "sha256": "244f40d77a40dfd184e78cf8e1e2dd9e469d3e594a8ce214279c2ce17b8f1012", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.title", "environments", "name", "scope", "summary", "topics"], "url": "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf"}], "id": "url:beb8f1fd6b945ffc9d93", "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:48+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html"}, {"rel": "publication_pdf", "url": "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf"}], "name": "GUI-SAGE: Enhancing GUI Automation with Self-Explanatory Learning", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. New work proposal from the official CVF accepted open-access version. No invented day, DOI, or arXiv identifier; original annotations describe the reviewed sources and not independent reproduction.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "GUI-SAGE generates reasoning conditioned on known actions and weights policy updates using confidence and reward signals. Its experiments use AndroidControl and GUIOdyssey data and separately evaluate interactive AndroidWorld tasks.", "topics": ["reinforcement-learning", "data-generation", "evaluation", "robustness"], "updated_at": null, "url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html"}
{"aliases": [], "authors": ["Allen Cypher"], "details": {"publication": {"authors": ["Allen Cypher"], "date_basis": "Supplied year precision: 1991. Opening CHI 1991 citation and copyright", "date_parts": [1991], "doi": null, "event_dates": null, "evidence_urls": ["https://acypher.com/Eager/", "https://acypher.com/Publications/CHI91/EagerCHI.html"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The 1991 conference paper is kept distinct from the 1991 video and 1993/1995 reprints listed on the author project page. Macintosh/HyperCard is tagged desktop; it is not represented as support for contemporary macOS. DOI not verified in a retained first-party source. Exact publication month/day is not inferred from conference dates.", "pages": "33-39", "pdf_url": null, "publisher": null, "series": null, "title": "Eager: Programming Repetitive Tasks by Example", "type": "conference-paper", "url": "https://acypher.com/Publications/CHI91/EagerCHI.html", "venue": "Proceedings of CHI '91", "volume": null}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:27Z", "role": "primary", "sha256": "64cffd3fca934465e1ea28bdf62e087bc9abdf50b9b72d04c2e57df0bcf41adb", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://acypher.com/Eager/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:21Z", "role": "primary", "sha256": "876d4c8dfac48443336a5ee97b463642db013631a2b97cacd090e61afd95ad59", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://acypher.com/Publications/CHI91/EagerCHI.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:21Z", "role": "primary", "sha256": "876d4c8dfac48443336a5ee97b463642db013631a2b97cacd090e61afd95ad59", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "scope", "summary", "topics"], "url": "https://acypher.com/Publications/CHI91/EagerCHI.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:27Z", "role": "primary", "sha256": "64cffd3fca934465e1ea28bdf62e087bc9abdf50b9b72d04c2e57df0bcf41adb", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.title", "name"], "url": "https://acypher.com/Eager/"}], "id": "url:bf0c11018e49dff6aeec", "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:55:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://acypher.com/Publications/CHI91/EagerCHI.html"}], "name": "Eager: Programming Repetitive Tasks by Example", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The 1991 conference paper is kept distinct from the 1991 video and 1993/1995 reprints listed on the author project page. Macintosh/HyperCard is tagged desktop; it is not represented as support for contemporary macOS.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Describes Eager, which detects repeated HyperCard interactions, generalizes them into loops, and previews predicted actions by highlighting interface objects. Users can then ask it to execute the inferred repetitive task.", "topics": ["automation", "human-in-the-loop", "interaction-history"], "updated_at": null, "url": "https://acypher.com/Publications/CHI91/EagerCHI.html"}
{"aliases": [], "authors": [], "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "84ce0fd682367a6dac5cc5abc39d5c877556024bbf6ffdad30f76fc2ec08523f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://hcompany.ai/research/models"}], "id": "url:c0bd07d82ce25ad1d99d", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://hcompany.ai/research/models"}], "name": "Holo Model Family", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://hcompany.ai/research/models", "target_id": "organization:h", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "H computer-use model family and its release-specific access options.", "topics": ["planning", "grounding"], "updated_at": null, "url": "https://hcompany.ai/research/models"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/showlab/ShowUI-2B/tree/main", "gated": false, "hub_id": "showlab/ShowUI-2B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "bb6d120d4795492a197e0a9e9df5c56b7baa8ac5c08182ce0f20f36c8b65783d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/showlab/ShowUI-2B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "bb6d120d4795492a197e0a9e9df5c56b7baa8ac5c08182ce0f20f36c8b65783d", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/showlab/ShowUI-2B"}], "id": "url:c0f39f7cf2cd9c18e8bd", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/showlab/ShowUI-2B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/showlab/ShowUI-2B"}], "name": "ShowUI 2B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/showlab/ShowUI-2B", "target_id": "github:showlab/showui", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "ShowUI vision-language-action checkpoint fine-tuned from Qwen2-VL-2B-Instruct, published as a single PyTorch weight file under the MIT license and tagged for GUI agents and computer use.", "topics": ["grounding", "efficiency"], "updated_at": null, "url": "https://huggingface.co/showlab/ShowUI-2B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "full recipe", "qualifier": "exact", "unit": "annotation rows", "value": 633369}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "point recipe", "qualifier": "exact", "unit": "annotation rows", "value": 610894}], "evidence_urls": ["https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"], "formats": ["JSONL", "JSON task recipes", "TSV manifests"], "modalities": ["text", "coordinates", "image references"], "notes": "The recipes reuse source images and are not additive datasets. This download contains no screenshots. Counts are publisher-reported; rows named for trajectories are independent decision examples, not full trajectories.", "splits": ["full/train", "point/train"], "supervision": ["instruction-to-click grounding", "bounding boxes", "absent-target abstention"]}, "license_name": "restricted-research-only"}, "environments": ["desktop", "web", "android", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "96a87222645c3f823e6f67e50708792e2917ddeb8eea5aacdf2de8c5bbff165f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "96a87222645c3f823e6f67e50708792e2917ddeb8eea5aacdf2de8c5bbff165f", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}], "id": "url:c1db2ba79e6d6a3db2b9", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations"}], "name": "ScreenRef Annotations", "notes": "The card declares restricted academic research use and no redistribution. Annotation coordinates use a normalized 1000 by 1000 space. Some source instructions name gestures other than the single-click target; the publisher recommends treating those as location supervision.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}, {"evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "target_id": "url:56bda04bd8d3a438deb3", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "Annotation files, task recipes, and image manifests for ScreenRef; screenshots reside in the separately gated image repository.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/api/datasets/osunlp/UGround-V1-Data", "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md"], "formats": ["Parquet shards"], "modalities": ["GUI screenshots", "natural-language grounding descriptions"], "notes": "Public API revision 33325877 lists Parquet assets but gated=auto; the pinned raw card returned HTTP 401 and no access was requested. Official README identifies this as V1 training data; the API excerpt also announces a Web-Hybrid bounding-box variant with coordinates normalized to [0,999]. That does not establish every shard schema or a release-wide point/bbox count. No train/test partition, example total, or unrestricted downloadability is claimed. Preserve the recorded CC-BY-NC-SA-4.0 data declaration separately from model/code terms.", "splits": [], "supervision": ["GUI visual-grounding training for the UGround-V1 family"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data/tree/main", "gated": "auto", "hub_id": "osunlp/UGround-V1-Data", "languages": [], "modalities": ["tabular", "text"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "0342862bd7857dd21e389ee5d351375dfea722cad4725c8e33271efc93e009f6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/osunlp/UGround-V1-Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "72ca3f23114c3a08fefb724573acf10c9f0a1b3d4a50d811c98356264ecf6ffc", "supports": ["details.dataset_profile", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "0342862bd7857dd21e389ee5d351375dfea722cad4725c8e33271efc93e009f6", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/osunlp/UGround-V1-Data"}], "id": "url:c4898ad1bfc049926a63", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "cc-by-nc-sa-4.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data"}], "name": "UGround V1 Data", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md", "target_id": "github:osu-nlp-group/uground", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "Gated web grounding training data for UGround in Web-Hybrid and Web-Direct subsets, with a bounding-box version whose coordinates are normalized to 0-999.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data"}
{"aliases": ["url:4a309ef39b33a7d1e543", "THUDM/cogagent-9b-20241220", "https://huggingface.co/THUDM/cogagent-9b-20241220"], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/zai-org/cogagent-9b-20241220/tree/main", "gated": false, "hub_id": "zai-org/cogagent-9b-20241220"}, "environments": ["desktop", "windows", "macos", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "de2c8e7a9792c3d129451ec813d458f83263460a5808cf87c45b04b260a376c9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:05Z", "role": "primary", "sha256": "ae72fbc40f9c8afd52d1688b2233e5c915571de33618cfd1ffcb7f3fadfb4402", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/cogagent-9b-20241220"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "ae72fbc40f9c8afd52d1688b2233e5c915571de33618cfd1ffcb7f3fadfb4402", "supports": ["identity"], "url": "https://huggingface.co/api/models/THUDM/cogagent-9b-20241220"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "c1ad1d99a8d46a00610db35d97c787d10b7e81661e48a1e389ae6cbb51a45630", "supports": ["licenses.weights"], "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/LICENSE"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "de2c8e7a9792c3d129451ec813d458f83263460a5808cf87c45b04b260a376c9", "supports": ["licenses.weights"], "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:13Z", "role": "primary", "sha256": "35bbe72f95711e042dab7a94bf2711e30a3da50190c7315178c37de5725cc1d9", "supports": ["aliases"], "url": "https://huggingface.co/api/models/THUDM/cogagent-9b-20241220"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:14Z", "role": "primary", "sha256": "35bbe72f95711e042dab7a94bf2711e30a3da50190c7315178c37de5725cc1d9", "supports": ["aliases"], "url": "https://huggingface.co/api/models/zai-org/cogagent-9b-20241220"}], "id": "url:c5eb975c0d66f8a2954f", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/cogagent-9b-20241220"}], "name": "CogAgent 9B 20241220", "notes": "The former THUDM model URL resolves to zai-org. Six public safetensors shards are present. Weights use the custom CogAgent License, including a registration requirement for commercial use, rather than a blanket permissive license. The card cites the 2023 CogAgent paper as previous work and links a later technical report; no claim that the old paper documents this 2024 checkpoint is asserted.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md", "target_id": "github:thudm/cogagent", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "A bilingual GLM-4V-based execution model that uses screenshots, task instructions and action history to predict GUI operations.", "topics": ["grounding", "action-prediction", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/zai-org/cogagent-9b-20241220"}
{"aliases": [], "authors": [], "details": {"availability": "Ungated benchmark JSONL/image manifest and public evaluation scripts; training pipeline/checkpoints are pending.", "benchmark_protocol": {"actions": ["critic-model responses"], "counts": [{"basis": "computed", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md", "population": "mobile/desktop/web configurations in the Hub card", "qualifier": "exact", "unit": "platform JSONL splits", "value": 3}], "evaluation": "The official repository runs model inference and computes accuracy with cal_acc.py.", "evidence_urls": ["https://raw.githubusercontent.com/numbmelon/OS-Oracle/HEAD/README.md", "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md"], "focus": ["GUI critic prediction"], "notes": "No sample count is inferred from the n<1K Hub size band. Evaluation code availability does not establish the release of OS-Oracle training data or weights.", "observations": ["benchmark JSONL", "UI screenshots"], "setting": "offline"}, "gated": false, "hub_id": "OS-Copilot/OS-Critic-Bench", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md"], "reason": "Publisher dataset-card license declaration for OS-Critic-Bench; not an evaluation-code or model-weight license."}}, "release_revision": "8fe774ef6a422c0e555f90205eb392f9132e272b"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:18Z", "role": "primary", "sha256": "58305b8ebe1af5dfbddd636484d63722d4e21d28ed3e22bb8c0068e654f33479", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/numbmelon/OS-Oracle/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:17Z", "role": "primary", "sha256": "17bcf851f39d06e7e357871df1dc1ddf697bcbea31842c16348227bbf4fbb224", "supports": ["identity"], "url": "https://api.github.com/repos/numbmelon/OS-Oracle"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:55Z", "role": "primary", "sha256": "516beea465382b54548e4b8bfd12ae510080a49451b30262ad1c769ed5a3398e", "supports": ["release_status", "licenses.code"], "url": "https://api.github.com/repos/numbmelon/OS-Oracle/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:06Z", "role": "primary", "sha256": "ef311b49437e246ffef6ad13c090277a0bdacc4a124e5b3cfe8d2643ced8f843", "supports": ["identity", "release_status", "details.hub_id", "details.gated", "details.release_revision"], "url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Critic-Bench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:07Z", "role": "primary", "sha256": "4139e25e8911ca8450f6db9acb8dc3a98c126c3c43514c72bcefa4ed33b8ffa8", "supports": ["licenses.data", "details.benchmark_protocol"], "url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md"}], "id": "url:c73ba0618d1b3c505887", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T11:18:18Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card license declaration for OS-Critic-Bench; not an evaluation-code or model-weight license.", "declared": "apache-2.0", "retrieved_at": "2026-10-17T11:23:07Z", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench"}], "name": "OS-Critic Bench", "notes": "The Hub card provides split/format/license metadata, while the official OS-Oracle README establishes critic-evaluation purpose and platform scope. The repository currently releases only benchmark inference/scoring components; data-synthesis code, training datasets and model checkpoints remain unchecked TODOs and are not counted as released assets. Apache-2.0 is recorded for benchmark data only; the evaluation code has no inspected license declaration.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/numbmelon/OS-Oracle/HEAD/README.md", "target_id": "arxiv:2512.16295", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A cross-platform GUI-critic benchmark with mobile, desktop and web JSONL splits and screenshot assets, accompanied by public inference and accuracy-scoring scripts.", "topics": ["evaluation", "reward-models", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "computed", "evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "population": "sum of icon/text counts in the published card table", "qualifier": "exact", "unit": "target annotations", "value": 1581}], "evidence_urls": ["https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py"], "formats": ["JSON annotations", "screen images"], "modalities": ["high-resolution screenshots", "instructions", "bounding boxes"], "notes": "Counts describe target annotations, not unique screenshots or expanded language/style variants. The evaluation code can reuse a target across configurations.", "splits": [], "supervision": ["GUI grounding"]}, "download_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/tree/main"}, "environments": ["desktop", "windows", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:35:55Z", "role": "primary", "sha256": "76ae2231564fe84af029f8625c02d9cd7fb6cf85c5f8af1b2a2593e2107719d9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:38:05Z", "role": "primary", "sha256": "a77c859775b33591b5619e8520c94702a53a1f9a29ada93e66744dbe23afe9dc", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:10Z", "role": "primary", "sha256": "2f7f635dee3c593b40efc61246c082ab1d76f7ce78f84a1d4b6391d30fa4e8fd", "supports": ["relations"], "url": "https://gui-agent.github.io/grounding-leaderboard/"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T11:35:55Z", "role": "primary", "sha256": "76ae2231564fe84af029f8625c02d9cd7fb6cf85c5f8af1b2a2593e2107719d9", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}], "id": "url:c79e152e6f7af1a9efed", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:35:55Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-16T11:35:55Z", "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro"}], "name": "ScreenSpot-Pro data", "notes": "This is the benchmark data asset, distinct from its evaluation-code repository. Target counts are not unique image counts.", "published_at": null, "relations": [{"evidence_url": "https://gui-agent.github.io/grounding-leaderboard/", "target_id": "arxiv:2504.07981", "type": "documented_by"}, {"evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "target_id": "github:likaixin2000/screenspot-pro-gui-grounding", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "High-resolution professional interface screenshots and target annotations for evaluating instruction-based GUI localization.", "topics": ["grounding", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/zonghanHZH/ZonUI-3B/tree/07212ea7724a69400422571e0f6e053212e66fe6", "gated": false, "hub_id": "zonghanHZH/ZonUI-3B", "hub_revision": "07212ea7724a69400422571e0f6e053212e66fe6", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md"], "reason": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate."}}, "repository_url": "https://github.com/Han1018/ZonUI-3B"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:16Z", "role": "primary", "sha256": "0dcd557508c07c9569e642a0319e71162b50c7ac6a0449564760ac2d6fffb089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.weights"], "url": "https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:22Z", "role": "primary", "sha256": "c419027b878ed7ebebd91f89a7395506d2f2aa8e1745918833cf1bac9e0a054e", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url"], "url": "https://huggingface.co/api/models/zonghanHZH/ZonUI-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "9918231748808ea1af93b559b0a511f1f9687689766a09b43c7726f14f27248b", "supports": ["release_status", "details.repository_url"], "url": "https://raw.githubusercontent.com/Han1018/ZonUI-3B/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:02Z", "role": "primary", "sha256": "7e3e63c1f2b38daaf4bb222b51fa5ae6594a0185e618723371e044bee6f05f7d", "supports": ["notes"], "url": "https://api.github.com/repos/Han1018/ZonUI-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:20Z", "role": "primary", "sha256": "d4c1500dbaceba738565d766e7b7453abbb6c0b12edaa34e5600d351214bfe4f", "supports": ["notes"], "url": "https://api.github.com/repos/Han1018/ZonUI-3B/git/trees/main?recursive=1"}], "id": "url:c8c046847f829fea4821", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-18T04:35:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:16Z", "url": "https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zonghanHZH/ZonUI-3B"}], "name": "ZonUI-3B checkpoint", "notes": "The Hub manifest lists four-part PyTorch and Safetensors files as alternative representations of one checkpoint, plus configuration/tokenizer files. The card’s example rescales predicted locations through the resized image dimensions. No checkpoint was loaded and no benchmark score was reproduced. Apache-2.0 is the model-card weight declaration; the separate code repository has no project-level license in the inspected tree.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md", "target_id": "arxiv:2506.23491", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Qwen2.5-VL-3B-based GUI grounding checkpoint that predicts click locations from screenshots and text instructions across desktop, mobile and web screens.", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": null, "url": "https://huggingface.co/zonghanHZH/ZonUI-3B"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B/tree/main", "gated": false, "hub_id": "zai-org/AutoGLM-Phone-9B", "languages": ["zh"], "modalities": [], "size_category_reported": null}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "94b02fa7104849cedee8ddd1fa33e22d3039e05ae9dd03033805a000e3fed847", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "94b02fa7104849cedee8ddd1fa33e22d3039e05ae9dd03033805a000e3fed847", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B"}], "id": "url:c918d5c828a3c6cfbc5d", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B"}], "name": "AutoGLM Phone 9B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B", "target_id": "github:zai-org/open-autoglm", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "AutoGLM phone-agent checkpoint for Open-AutoGLM, fine-tuned from GLM-4.1V-9B-Base with about 10.3B BF16 parameters and tagged for Chinese.", "topics": ["mobile-navigation"], "updated_at": null, "url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md", "population": "publisher-reported MobileViews-600K first release", "qualifier": "more-than", "unit": "screenshot–view-hierarchy pairs", "value": 600000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md", "population": "publisher-reported MobileViews-600K first release", "qualifier": "more-than", "unit": "Android apps", "value": 20000}], "evidence_urls": ["https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md", "https://huggingface.co/api/datasets/mllmTeam/MobileViews"], "formats": ["JPEG", "JSON", "XML", "Parquet", "CSV", "ZIP packages", "HTML/JavaScript trace visualization"], "modalities": ["screenshots", "view hierarchies", "action traces", "app metadata"], "notes": "Counts describe the card’s first-release claim, not an independently counted archive inventory. The card distinguishes globally image-hash-deduplicated screenshot/VH pairs from the November 2024 complete-app traces; trace state counts must not be added as new unique screens. JPEG/JSON and Parquet are alternative representations. Listed filename ranges are examples, not a complete count proof. No train/validation/test split is documented. Hub revision c734715e is ungated; MIT is its data declaration, without resolving third-party app content rights.", "splits": [], "supervision": ["paired screenshot/view-hierarchy structure", "recorded app states, actions and transitions"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/mllmTeam/MobileViews/tree/main", "gated": false, "hub_id": "mllmTeam/MobileViews", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "d36d62283deecd3c5ccf1ac1af49aa7ba779a1d69cdcaea39a1bed37ff5d66b9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/mllmTeam/MobileViews"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:10:01Z", "role": "primary", "sha256": "94a4f04e180dbd597c271b0d7470c907e6526b95a6420f01b3183feb1ee24d6e", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "d36d62283deecd3c5ccf1ac1af49aa7ba779a1d69cdcaea39a1bed37ff5d66b9", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/mllmTeam/MobileViews"}], "id": "url:cc4aa8caeee0cff58e73", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/mllmTeam/MobileViews"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/mllmTeam/MobileViews"}], "name": "MobileViews", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/mllmTeam/MobileViews", "target_id": "arxiv:2409.14337", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Over 600,000 Android screenshot and view-hierarchy pairs from more than 20,000 Google Play apps, collected with an optimized DroidBot, with complete per-app traces.", "topics": ["screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/mllmTeam/MobileViews"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/tencent/POINTS-GUI-G/tree/main", "gated": false, "hub_id": "tencent/POINTS-GUI-G"}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "ac679d3cfdda8754e945d127935c9ed9030e6c9d0c9872f6224909ab2abe5c3f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "3b97405c0a7f50298cf31a4933d31da356363c2ba5ecfaf73144cedeebc44db2", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/tencent/POINTS-GUI-G"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "9f851237932688c15d261c016d51ff5ae80c43c5d0b4eaccac0b48d941118d80", "supports": ["environments"], "url": "https://raw.githubusercontent.com/Tencent/POINTS-GUI/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "ac679d3cfdda8754e945d127935c9ed9030e6c9d0c9872f6224909ab2abe5c3f", "supports": ["licenses.weights"], "url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md"}], "id": "url:cc931199bda2bdf1f5f7", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/tencent/POINTS-GUI-G"}], "name": "POINTS-GUI-G 8B", "notes": "Four safetensors shards and model code/configuration are public and ungated. The card declares license other without specifying a named weight license. This release is the grounding model; the same card says the end-to-end GUI agent model is still under development. No end-to-end agent capability is inferred from grounding scores.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md", "target_id": "arxiv:2602.06391", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A specialized interface-grounding checkpoint that returns normalized point coordinates for a requested element in a screenshot.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/tencent/POINTS-GUI-G"}
{"aliases": [], "authors": ["Greg Little", "Tessa A. Lau", "Allen Cypher", "James Lin", "Eben M. Haber", "Eser Kandogan"], "details": {"publication": {"authors": ["Greg Little", "Tessa A. Lau", "Allen Cypher", "James Lin", "Eben M. Haber", "Eser Kandogan"], "date_basis": "Supplied year precision: 2007. CHI 2007 proceedings heading and copyright year The IBM page has an unlabeled Date of 22 Oct 2007, while the paper gives April/May event dates. Their roles do not establish a reliable day-level publication date.", "date_parts": [2007], "doi": "10.1145/1240624.1240767", "event_dates": null, "evidence_urls": ["https://research.ibm.com/publications/koala-capture-share-automate-personalize-business-processes-on-the-web", "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "This proposal is the specific Koala paper. It does not automatically merge every later CoScripter publication or project record. Exact publication date is unresolved; only the clearly established year is proposed.", "pages": "943-946", "pdf_url": "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf", "publisher": null, "series": null, "title": "Koala: Capture, Share, Automate, Personalize Business Processes on the Web", "type": "conference-paper", "url": "https://doi.org/10.1145/1240624.1240767", "venue": "CHI 2007 Proceedings", "volume": null}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:23Z", "role": "primary", "sha256": "ae61a94cca68e058858244b6ff54ed9409973a5cecaf25c0e325b27adaf0e3dd", "supports": ["details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://research.ibm.com/publications/koala-capture-share-automate-personalize-business-processes-on-the-web"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:21Z", "role": "primary", "sha256": "944b6f74a6cdb88fc253009f364dcc71cc42186b4907e04720da128c5292e357", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.venue"], "url": "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:21Z", "role": "primary", "sha256": "944b6f74a6cdb88fc253009f364dcc71cc42186b4907e04720da128c5292e357", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:23Z", "role": "primary", "sha256": "ae61a94cca68e058858244b6ff54ed9409973a5cecaf25c0e325b27adaf0e3dd", "supports": ["details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://research.ibm.com/publications/koala-capture-share-automate-personalize-business-processes-on-the-web"}], "id": "url:cccd967aa993404d74d0", "identifiers": {"doi": "10.1145/1240624.1240767"}, "kind": "paper", "last_verified_at": "2026-10-17T09:55:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/1240624.1240767"}, {"rel": "publication_pdf", "url": "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf"}], "name": "Koala: Capture, Share, Automate, Personalize Business Processes on the Web", "notes": "The annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. This proposal is the specific Koala paper. It does not automatically merge every later CoScripter publication or project record.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Introduces Koala, a collaborative web automation system that records browser interactions as editable pseudo-natural-language scripts. Scripts are shared through a wiki and instantiated with personal or corporate data for reuse by other users.", "topics": ["web-navigation", "workflow", "recording"], "updated_at": null, "url": "https://doi.org/10.1145/1240624.1240767"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/microsoft/Phi-Ground/tree/main", "gated": false, "hub_id": "microsoft/Phi-Ground", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "ce8e0ebc3d4c2642bfb6a1d807176eae4c14b52fbc6f221b3c234a9d55e0d362", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/microsoft/Phi-Ground"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "ce8e0ebc3d4c2642bfb6a1d807176eae4c14b52fbc6f221b3c234a9d55e0d362", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/Phi-Ground"}], "id": "url:cee45b940d5fbff3ccca", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/microsoft/Phi-Ground"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/Phi-Ground"}], "name": "Phi-Ground checkpoint", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/microsoft/Phi-Ground", "target_id": "github:microsoft/phi-ground", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "Microsoft Phi-Ground GUI grounding checkpoint fine-tuned from Phi-3.5-vision-instruct, shipped as two PyTorch weight shards with custom Phi-3-V configuration and processing code under MIT.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/microsoft/Phi-Ground"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md", "population": "AndroidCode corpus described by Code2World README; released JSON rows not recounted", "qualifier": "more-than", "unit": "screen–action pairs", "value": 80000}], "evidence_urls": ["https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md", "https://huggingface.co/api/datasets/GD-ML/AndroidCode"], "formats": ["HTML", "JSON", "Python extraction/loading scripts"], "modalities": ["GUI screenshots (described input)", "GUI actions", "HTML"], "notes": "The manifest lists 58,953 HTML files, which are not converted into transition/example counts. The reported >80k pairs and the HTML-file inventory have different units. The card itself has no sample schema or count; archive/image completeness and JSON contents were not audited.", "splits": ["Android-Control train file", "Android-Control test file"], "supervision": ["action-conditioned next-state code prediction", "generated HTML supervision"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/GD-ML/AndroidCode/tree/af316a6be79bbe08c4fc09ea588b5f80f177ed25", "gated": false, "hub_id": "GD-ML/AndroidCode", "hub_revision": "af316a6be79bbe08c4fc09ea588b5f80f177ed25", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/GD-ML/AndroidCode/raw/af316a6be79bbe08c4fc09ea588b5f80f177ed25/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}, "repository_url": "https://github.com/AMAP-ML/Code2World"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "26c14a14166d3b747454d1924ecdf6d96652ac857ae6ad488630dc843036a777", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:22Z", "role": "primary", "sha256": "5774a3c096264fbba58b7a16587ea65f35628edb9ca39afe8a32c74ec7fd1ce7", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/GD-ML/AndroidCode"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:17Z", "role": "primary", "sha256": "d8d7a46d41a1a37fe4f0a5f637bf55c649310185329127d8a2204632e480be17", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/GD-ML/AndroidCode/raw/af316a6be79bbe08c4fc09ea588b5f80f177ed25/README.md"}], "id": "url:cf4d2d45a7314cb16acb", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:27:14Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:17Z", "url": "https://huggingface.co/datasets/GD-ML/AndroidCode/raw/af316a6be79bbe08c4fc09ea588b5f80f177ed25/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/GD-ML/AndroidCode"}], "name": "AndroidCode", "notes": "The Hub card is license-only; semantic description and reported scale come from the official Code2World README. The public manifest contains HTML files, Android-Control train/test JSON files and extraction/loading scripts. It does not list standalone screenshot image files, and embedded/reference image packaging was not inspected. MIT is the data-card declaration, not a grant for every upstream source dataset.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md", "target_id": "arxiv:2602.09856", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "HTML and annotation release used for Code2World’s action-conditioned GUI-state prediction training, constructed from mobile GUI trajectories.", "topics": ["world-models", "data-generation", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/GD-ML/AndroidCode"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/cckevinn/SeeClick/tree/main", "gated": false, "hub_id": "cckevinn/SeeClick", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["web", "desktop", "android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:52Z", "role": "primary", "sha256": "bc9f7c36f3e7c4515fa4982b0dd921a9038ff082cb6b283572d6390ef651c2a9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/cckevinn/SeeClick"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "c4112415bec58fc77f9a1e8fa98f8d602c5d382747579dfda9f866900c55995c", "supports": ["summary", "environments", "relations", "licenses"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:05:52Z", "role": "primary", "sha256": "bc9f7c36f3e7c4515fa4982b0dd921a9038ff082cb6b283572d6390ef651c2a9", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/cckevinn/SeeClick"}], "id": "url:cfea8a0949472a637990", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:05:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/cckevinn/SeeClick"}], "name": "SeeClick checkpoint", "notes": "The public ungated Hub manifest contains ten safetensors shards and model configuration/code but no README model card or explicit weight-license field. The official repository identifies the checkpoint and says incorporated checkpoints follow their original licenses, without resolving a specific standalone weight license. No source or model code was executed.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md", "target_id": "arxiv:2401.10935", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The released SeeClick visual grounding checkpoint for mapping an interface screenshot and referring instruction to a point or bounding box.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/cckevinn/SeeClick"}
{"aliases": [], "authors": [], "details": {"availability": "Ungated JSONL artifact manifest; no dataset card or license declaration.", "dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md", "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/data/dataset.md", "https://huggingface.co/api/datasets/wuuuuuz/GUI-CIDER"], "formats": ["JSONL"], "modalities": ["text"], "notes": "Purpose is established by the official project, not by API tags. No sample records were downloaded; count and data-license fields remain unknown.", "splits": [], "supervision": ["synthesized static planning knowledge", "synthesized dynamic causal knowledge"]}, "gated": false, "hub_id": "wuuuuuz/GUI-CIDER", "release_revision": "47ba5a55ed9ad993ad9562904ea6e978c4524d17"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:20Z", "role": "primary", "sha256": "9a46741bbde1373ed4b87887cdadde98e05ea8795d34efe39fc25d69b4d50213", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:07Z", "role": "primary", "sha256": "aec6da0e872092d737953f8b4db227867024da10e127cd32d15d05ff95552d45", "supports": ["identity", "release_status", "details.hub_id", "details.gated", "details.release_revision"], "url": "https://huggingface.co/api/datasets/wuuuuuz/GUI-CIDER"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:24:47Z", "role": "primary", "sha256": "bffca456e649a08239f6a1a3be10a7d7c21bcb427e783ba2c304e4b8410d385a", "supports": ["environments", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/data/dataset.md"}], "id": "url:d0300db5ba754b1a64fa", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T11:18:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/wuuuuuz/GUI-CIDER"}], "name": "GUI-CIDER causal knowledge data", "notes": "The official project explicitly links this corpus, and the ungated Hub API lists GUI-CIDER.jsonl. The repository has no README card at the Hub (404); purpose and Android context are therefore supported by the project README and dataset guide, while the API supports only identity/file availability. No sample count, split inventory, or data license was established. The project README’s MIT badge is not transferred to this data artifact.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md", "target_id": "github:wuzheng02/gui-cider", "type": "component_of"}, {"evidence_url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md", "target_id": "arxiv:2605.28534", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A published JSONL corpus of synthesized planning and causal knowledge for GUI-agent mid-training, linked directly from the official GUI-CIDER project.", "topics": ["data-generation", "imitation-learning", "planning"], "updated_at": null, "url": "https://huggingface.co/datasets/wuuuuuz/GUI-CIDER"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI/tree/main", "gated": false, "hub_id": "Miaosen/Grounding-Anything-GUI", "languages": ["en"], "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "c64f72eae3dd50d28996f99ab744a5093896e8bb29caa6f7b33a2f60a0b530a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/Miaosen/Grounding-Anything-GUI"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "31b964cfd198f99491f0f480024b2322fbdf3a237f7806847e83978109bc5856", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a0609394a736dc495d077cd57a2945b018123e5b3f9514d7d5bafb9e690aeae2", "supports": ["summary", "release_status", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "c64f72eae3dd50d28996f99ab744a5093896e8bb29caa6f7b33a2f60a0b530a5", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Miaosen/Grounding-Anything-GUI"}], "id": "url:d1830de02226155ee16d", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:42:32Z", "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI"}], "name": "Grounding-Anything-GUI", "notes": "The Hub exposes train.zip and a card containing only license/language metadata. The official repository marks this training-data release complete, but its Phi-Ground-Any paper is still pending. No binding to the earlier Phi-Ground paper, sample count, split structure or operating-system coverage is inferred.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md", "target_id": "github:microsoft/phi-ground", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "The public training archive linked by the Phi-Ground project for its Phi-Ground-Any model.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/zai-org/webrl-orm-llama-3.1-8b/tree/main", "gated": false, "hub_id": "zai-org/webrl-orm-llama-3.1-8b", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "8bafcdf63fe484b6005ca7082da1310459f8c87d4c011e84293a3cc8fca6ae9e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/zai-org/webrl-orm-llama-3.1-8b"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:29Z", "role": "primary", "sha256": "8bafcdf63fe484b6005ca7082da1310459f8c87d4c011e84293a3cc8fca6ae9e", "supports": ["identity"], "url": "https://huggingface.co/api/models/THUDM/webrl-orm-llama-3.1-8b"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "eb217b96c54c3c2ccfd3a27255c6fd740df40c2d6743ae811858339bf7f740b0", "supports": ["summary", "environments", "relations"], "url": "https://raw.githubusercontent.com/THUDM/WebRL/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "8bafcdf63fe484b6005ca7082da1310459f8c87d4c011e84293a3cc8fca6ae9e", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/webrl-orm-llama-3.1-8b"}], "id": "url:d1e85aeb6d96859c109d", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:05:40Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/webrl-orm-llama-3.1-8b"}], "name": "WebRL outcome reward model 8B", "notes": "The official API resolves the former THUDM namespace to zai-org and lists four safetensors shards. No model card or weight-license declaration is present in the inspected manifest. The official WebRL README identifies this as its outcome-supervised reward model; it is a training component rather than an action policy.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/WebRL/HEAD/README.md", "target_id": "arxiv:2411.02337", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "The outcome-supervision checkpoint used by WebRL to score browser trajectories during policy training.", "topics": ["reward-models", "reinforcement-learning", "web-navigation"], "updated_at": null, "url": "https://huggingface.co/zai-org/webrl-orm-llama-3.1-8b"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual/tree/main", "gated": false, "hub_id": "zai-org/AutoGLM-Phone-9B-Multilingual", "languages": ["zh"], "modalities": [], "size_category_reported": null}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "d66bc86d9fb61f88d1456b36cb0c2c9ab5e178b55bc9d34f307b35733f9b90bc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B-Multilingual"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "d66bc86d9fb61f88d1456b36cb0c2c9ab5e178b55bc9d34f307b35733f9b90bc", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B-Multilingual"}], "id": "url:d236e7ec7a2b0028950a", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual"}], "name": "AutoGLM Phone 9B Multilingual", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B-Multilingual", "target_id": "github:zai-org/open-autoglm", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "Multilingual variant of the AutoGLM phone-agent checkpoint for Open-AutoGLM, likewise fine-tuned from GLM-4.1V-9B-Base with about 10.3B BF16 parameters.", "topics": ["mobile-navigation"], "updated_at": null, "url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "832b84c366eaadc69e328f207f2659cf07a0901c47662244831c0f1695d02efd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://manus.im/"}], "id": "url:d3414794302313051ff0", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://manus.im/"}], "name": "Manus", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://manus.im/", "target_id": "organization:manus", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "General agent platform with browser-based task execution.", "topics": ["workflow"], "updated_at": null, "url": "https://manus.im/"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-9B", "gated": false, "hub_revision": "68ebbac4d5bdf725103a22d076122b8aa883a0f2", "listed_weight_files": ["model.safetensors"]}, "environments": ["android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "09c59e6b35a5fe1eea528595f326546269ea8a2a75aa404f1f7950b9de5f30d1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "1aab6befd309e54ba71676670039b4098362974ce0ce89b25872f137d50c103f", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/OpenMobile-2/OpenMobile-2-9B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:06Z", "role": "primary", "sha256": "9b385e9f6aa2f7169f616a76887df3b76df605cb2bea071d99ee7ca3f9b659cc", "supports": ["environments"], "url": "https://os-copilot.github.io/OpenMobile2-Home/"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "09c59e6b35a5fe1eea528595f326546269ea8a2a75aa404f1f7950b9de5f30d1", "supports": ["licenses.weights"], "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md"}], "id": "url:d35224f787e7ec21ce36", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:51Z", "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B"}], "name": "OpenMobile-2 9B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md", "target_id": "github:openmobile2/openmobile2-code", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md", "target_id": "url:546e6c95825d1318c98d", "type": "trained_on"}, {"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md", "target_id": "arxiv:2610.14157", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile-agent checkpoint trained on OpenMobile-Data-v2 for screenshot actions and app-native tool calls.", "topics": ["mobile-navigation", "tool-integration"], "updated_at": null, "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md", "https://huggingface.co/api/datasets/OS-Copilot/OS-Genesis-mobile-data", "https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md"], "formats": ["JSONL", "ZIP"], "modalities": ["screenshots", "text", "actions"], "notes": "AC and AW refer to distinct target training sets; no sample count is inferred from the two JSONL files or ZIP packaging.", "splits": ["AndroidControl training", "AndroidWorld training"], "supervision": ["mobile action trajectories"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Genesis-mobile-data"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "de9de63e21e60e882277a6268e8153abb6c14aaf3e9a7f83e7f98ac14bf37170", "supports": ["name", "kind", "relations", "licenses.data"], "url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "6a5e0c2baecc2dc8bf4466903798f5ecda07cdac917042bbd4cfbe54d895c11e", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Genesis-mobile-data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "430e495ab250395d666854bd18b0fc29b819cd784cd7730215ef65e461834d0d", "supports": ["summary", "environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "de9de63e21e60e882277a6268e8153abb6c14aaf3e9a7f83e7f98ac14bf37170", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md"}], "id": "url:d3ec5122a2b52fd5ae4b", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:34Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:01:34Z", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data"}], "name": "OS-Genesis mobile training data", "notes": "The public manifest contains AC/AW JSONL training annotations and a screenshot archive. MIT is declared for the data in its own card, separately from Apache-2.0 declarations on some model checkpoints. The gated web-data release is not included in this asset.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md", "target_id": "arxiv:2412.19723", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Reverse-task-synthesis mobile trajectories released in separate AndroidControl and AndroidWorld training annotation files.", "topics": ["data-generation", "mobile-navigation", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["success verdict", "process score"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md", "population": "fara7b_om2w_browserbase split", "qualifier": "exact", "unit": "trajectories", "value": 106}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md", "population": "internal split", "qualifier": "exact", "unit": "trajectories", "value": 154}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md", "population": "fara7b_om2w_browserbase split", "qualifier": "exact", "unit": "annotation rows", "value": 215}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md", "population": "internal split", "qualifier": "exact", "unit": "annotation rows", "value": 154}], "evaluation": "Agreement metrics including Cohen’s kappa, accuracy and F1 against human annotations; review stages must be kept distinct.", "evidence_urls": ["https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md"], "focus": ["verifier-human agreement", "process quality"], "notes": "Trajectory rows and reviewer rows have different grains. Internal task labels do not make the released Parquet files private. No payload data was downloaded.", "observations": ["screenshots", "action logs", "final answers"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/tree/main", "gated": false, "hub_id": "microsoft/CUAVerifierBench", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md"], "reason": "The benchmark's Hub dataset card declares MIT in its metadata header (license: mit)."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "44d7190b72ae489297f59d6fadff635ced6dd067c845c77b04f187a7b2d7bec7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "fb1fe8a9fd6e45fc82157c65c08d5c682d08bead5196308776200ef3c39c1e41", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/microsoft/CUAVerifierBench"}], "id": "url:d5140d438466c1e1348a", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:42:32Z", "url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench"}], "name": "CUAVerifierBench", "notes": "Public Parquet configs separate trajectories from per-reviewer annotations. The card declares MIT. The two splits have different review stages: Online-Mind2Web includes blind and verifier-informed annotations, whereas the internal split is blind-only. This measures verifier agreement rather than agent task success.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md", "target_id": "arxiv:2604.06240", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Human-annotated browser-agent trajectories for comparing automated task-success and process-quality verifiers against reviewer judgments.", "topics": ["evaluation", "reward-models", "web-navigation"], "updated_at": null, "url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/xuyifan/MobileRL-9B/tree/main", "gated": false, "hub_id": "xuyifan/MobileRL-9B", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "a00f13f5a86f2de7eb9c866cc4d08b958d21b2ac30dc3b9e1662e4bb336b7cb3", "supports": ["name", "kind", "relations", "licenses.weights", "identity", "release_status", "details"], "url": "https://huggingface.co/api/models/xuyifan/MobileRL-9B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "5c2d5eb247958f1af0646cbac831ecfbc7f839baa236a0d733075592da81f58e", "supports": ["licenses.weights"], "url": "https://huggingface.co/xuyifan/MobileRL-9B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "752fad8ebd8b8eda03648329a6fc2585e5bcb2b96875b8aa2ce7d9f3fff78153", "supports": ["summary", "environments", "relations", "topics", "scope"], "url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "a00f13f5a86f2de7eb9c866cc4d08b958d21b2ac30dc3b9e1662e4bb336b7cb3", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/xuyifan/MobileRL-9B"}], "id": "url:d5667c854845c77b9f5b", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:40Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:03:40Z", "url": "https://huggingface.co/xuyifan/MobileRL-9B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/xuyifan/MobileRL-9B"}], "name": "MobileRL 9B checkpoint", "notes": "Five public safetensors shards are present. The short card declares Apache-2.0 and a GLM-4.1V-9B base. The official project confirms the checkpoint and Android targets, and distinguishes MobileRL bounding-box actions from coordinate-format alternatives.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/README.md", "target_id": "arxiv:2509.18119", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A GLM-4.1V-based mobile action checkpoint released by the MobileRL project for Android interaction.", "topics": ["reinforcement-learning", "mobile-navigation"], "updated_at": null, "url": "https://huggingface.co/xuyifan/MobileRL-9B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/tree/main", "gated": false, "hub_id": "InfiX-ai/InfiGUI-G1-7B"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "253055f0f77266a64d2db5de8fd04c4202f4cda3be733dd1b00d8565116dd240", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "10278eb0f1600557e432a0f8dde2c40cfffe12330ddb949b6a0fc02bfbc6c9ee", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/InfiX-ai/InfiGUI-G1-7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "4b68c82a2dd0063a450bde34c8bffbd931565a91a2982a5b47be7e711815b284", "supports": ["environments"], "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "253055f0f77266a64d2db5de8fd04c4202f4cda3be733dd1b00d8565116dd240", "supports": ["licenses.weights"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md"}], "id": "url:d70a755e3b2102f25f34", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B"}], "name": "InfiGUI-G1 7B", "notes": "Four safetensors shards are public. The model card declares Apache-2.0 and identifies Qwen2.5-VL-7B-Instruct as its base. It is the G1 grounding checkpoint, distinct from the previously cataloged InfiGUI-R1 reasoning checkpoint; benchmark scores remain publisher claims.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md", "target_id": "arxiv:2508.05731", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A GUI-grounding checkpoint trained with adaptive exploration rewards and multiple candidate answers to improve instruction-to-element alignment.", "topics": ["grounding", "reinforcement-learning"], "updated_at": null, "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B"}
{"aliases": [], "authors": [], "details": {"benchmark_protocol": {"actions": ["mobile interface operations", "clarification requests"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md", "population": "MobileGUI-VBench release", "qualifier": "exact", "unit": "tasks", "value": 150}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md", "population": "MobileGUI-VBench release", "qualifier": "exact", "unit": "mobile apps", "value": 40}], "evaluation": "Published task definitions include preconditions and clarification requirements; the metadata review does not independently validate the reported agent scores.", "evidence_urls": ["https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md"], "focus": ["mobile task completion", "clarification behavior", "cross-app tasks"], "notes": "The card describes 140 single-app and 10 cross-app tasks. Task and app counts are different units.", "observations": ["mobile app interfaces"], "setting": "interactive"}, "disabled": false, "download_url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/tree/main", "gated": false, "hub_id": "vivo-ai/MobileGUI-VBench"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "422ff6a557a1aabd8f6c6ef26e5fd05b992f4acf0e5128ab8bc437021fcc5ea7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "5850fc58f54c877cffe328dba2a0b5e3e98767637aae6d6efd9a037cf06eabce", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/vivo-ai/MobileGUI-VBench"}], "id": "url:d7a4f34a6a5202af5f64", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench"}], "name": "MobileGUI-VBench", "notes": "Public task files are available as JSONL and XLSX with documentation. The card explicitly declares Apache-2.0 for the dataset, despite having no license field in its YAML metadata. It identifies mobile apps without fixing an operating-system taxonomy; model scores are not imported.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md", "target_id": "arxiv:2609.12394", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Mobile-app task definitions covering everyday scenarios, with annotated preconditions and clarification requirements for evaluating agent behavior.", "topics": ["evaluation", "mobile-navigation", "cross-app"], "updated_at": null, "url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview/tree/main", "gated": false, "hub_id": "stepfun-ai/GELab-Zero-4B-preview", "languages": ["en", "zh"], "modalities": [], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "cd1efa947316f13967165ac906400170df6d28446804ba0763a49fc21a731e7e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/stepfun-ai/GELab-Zero-4B-preview"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "cd1efa947316f13967165ac906400170df6d28446804ba0763a49fc21a731e7e", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/stepfun-ai/GELab-Zero-4B-preview"}], "id": "url:d7ad447ecd712c6d51b8", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview"}], "name": "GELab Zero 4B Preview", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/stepfun-ai/GELab-Zero-4B-preview", "target_id": "github:stepfun-ai/gelab-zero", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "Preview GELab-Zero phone-use agent checkpoint fine-tuned from Qwen3-VL-4B-Instruct, with about 4.4B BF16 parameters, tagged for Android in English and Chinese, under Apache-2.0.", "topics": ["mobile-navigation"], "updated_at": null, "url": "https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "461de0060fae1ded024dfa724624b001665b463b14f454afafebd0c9a2dce71d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.tinyfish.ai/"}], "id": "url:d92596cfe618e5322ec2", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.tinyfish.ai/"}], "name": "TinyFish", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://www.tinyfish.ai/", "target_id": "organization:tinyfish", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Browser-agent API for executing web tasks and extracting structured results.", "topics": ["web-navigation"], "updated_at": null, "url": "https://www.tinyfish.ai/"}
{"aliases": [], "authors": [], "details": {"leaderboard_url": "https://recreation-bench.cc/#leaderboard"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "79deaa899d085ece77a127d53b12ddb9edea8a796943f35b75e9acec67bcb976", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/Qwen/RecreationBench/raw/main/README.md"}], "id": "url:d9d8e68f19ef0956d189", "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Qwen/RecreationBench"}], "name": "RecreationBench", "notes": "Dataset card declares multiple upstream licenses; do not treat the package as having one blanket permissive license.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Tasks for recreating reference applications with hybrid computer interaction.", "topics": ["evaluation", "tool-integration"], "updated_at": null, "url": "https://huggingface.co/datasets/Qwen/RecreationBench"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B/tree/main", "gated": false, "hub_id": "apurvaga/go-browse-wa-qwen-7B"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "470ef6f975990518f9622b259a6b8e5761b26dcd22985c248f1df2beb4aa2760", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "d083bf53b20c6c6124f1df29a83ce783799d46a1ab5600d19d658729fc6d67bb", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/apurvaga/go-browse-wa-qwen-7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "0d4b28c993ccb977c7910ff3fecb5de29db364a7a7ad55523c8e052c9bae2566", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md"}], "id": "url:db1b7b7efe3755ae5693", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B"}], "name": "Go-Browse WebArena Qwen 7B", "notes": "Seven public safetensors shards and model/tokenizer files are present. The card links the official code and paper but declares no weight license. The code repository’s MIT license is not inherited by this checkpoint, and the separate NNetNav-trained comparison model is not treated as an alias.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B/raw/main/README.md", "target_id": "arxiv:2506.03533", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen-2.5-7B-Instruct checkpoint fine-tuned on Go-Browse WebArena trajectories for website task execution.", "topics": ["web-navigation", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B/tree/2c5604d86c529843ef346698edcf33b170c1e817", "gated": false, "hub_id": "Fanbin/ARPO_UITARS1.5_7B", "hub_revision": "2c5604d86c529843ef346698edcf33b170c1e817", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B/raw/2c5604d86c529843ef346698edcf33b170c1e817/README.md"], "reason": "The inspected model card contains no checkpoint-license declaration. The code repository Apache license is not transferred to weights."}}, "repository_url": "https://github.com/JIA-Lab-research/ARPO"}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:12Z", "role": "primary", "sha256": "839b2de0e0c614fe8d2df4782d0f6a6e7d33f5c60d8020c9d4a01c61d533f39e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:21Z", "role": "primary", "sha256": "6e19a5339c4e52ca3753cce26ac2034883b96f5cb26383a4b41bc87a5fbf5bf6", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url"], "url": "https://huggingface.co/api/models/Fanbin/ARPO_UITARS1.5_7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:16Z", "role": "primary", "sha256": "b6957fad37a7995e9dde4c7d4a4e0f6e2153b85354b2d234587c2a94ff95d0bd", "supports": ["summary", "notes", "licenses.weights"], "url": "https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B/raw/2c5604d86c529843ef346698edcf33b170c1e817/README.md"}], "id": "url:db6fc5f29f186d13991c", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-18T04:27:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B"}], "name": "ARPO UI-TARS-1.5-7B checkpoint", "notes": "The public Hub manifest lists four Safetensors shards, an index and configuration/tokenizer files. The model card reports a 15-step evaluation setting; no performance result is independently reproduced. No weight license is declared, despite Apache-2.0 being established for the separate code repository.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/README.md", "target_id": "arxiv:2505.16282", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "UI-TARS-1.5-7B-derived checkpoint trained with ARPO on OSWorld desktop tasks.", "topics": ["reinforcement-learning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3-VL-8B-Instruct", "gated": false, "hub_revision": "9db378189fddaabb5f6e347b9053858a392022ac", "listed_weight_files": ["model.safetensors"]}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "73a4690fab49d3b08ea9b0b1e44d62d5ec1802614f5cf22c37ed92dca49aa520", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "8b680b9a63e654fe0b43754201d30c91714859578f294f4781e035280e862657", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/PrentisAI/GroundGUI-8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "73a4690fab49d3b08ea9b0b1e44d62d5ec1802614f5cf22c37ed92dca49aa520", "supports": ["licenses.weights"], "url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md"}], "id": "url:de31b07ff24ab15d9bdf", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/PrentisAI/GroundGUI-8B"}], "name": "GroundGUI-8B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md", "target_id": "github:prentisai/groundgui", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md", "target_id": "url:56bda04bd8d3a438deb3", "type": "trained_on"}, {"evidence_url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Screenshot grounding checkpoint that emits a single click in a normalized coordinate space.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/PrentisAI/GroundGUI-8B"}
{"aliases": ["Nova Act"], "authors": [], "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:18:21Z", "role": "primary", "sha256": "4e4e60092351a85d0b99186a745d6295b170958be59092a786eda369923bd691", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://aws.amazon.com/nova/act/"}], "id": "url:dec156cbf0cb1181f7b6", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-17T17:18:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://aws.amazon.com/nova/act/"}], "name": "Amazon Nova Act", "notes": "", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "AWS service for building and deploying agents that automate browser-based UI workflows from natural language and Python, powered by a custom Nova 2 Lite model.", "topics": ["automation", "web-navigation"], "updated_at": null, "url": "https://aws.amazon.com/nova/act/"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/ritzzai/GUI-R1/tree/main", "gated": false, "hub_id": "ritzzai/GUI-R1"}, "environments": ["desktop", "windows", "linux", "macos", "android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "f31ae9547d94967dd91f5cde5e5c66430d7d59ee3a525c0914bc600a79999af7", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/ritzzai/GUI-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "94c7c78bede11dde727940f56592a04594c1c59438812dff47540bf53c1bed7d", "supports": ["summary", "environments", "release_status"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md"}], "id": "url:df680385f18b01a825be", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:42:32Z", "url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/ritzzai/GUI-R1"}], "name": "GUI-R1 data bundle", "notes": "The Hub is a dataset repository with Parquet payloads, although the short card mistakenly calls it a model. The official project identifies it as training/evaluation data and describes subsequent expanded mixtures. MIT is declared in this dataset card; upstream dataset terms remain separate. No 3K/800K aggregate count is inferred for the mutable bundle.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md", "target_id": "arxiv:2504.10458", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Training and evaluation Parquet bundles accompanying GUI-R1, including GUI grounding and mobile/web action-prediction data.", "topics": ["grounding", "reinforcement-learning", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/ritzzai/GUI-R1"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xlangai/Jedi/raw/main/README.md"], "formats": [], "modalities": ["images", "text", "coordinates"], "notes": "Coordinates were resized for a specified Qwen2.5-VL image processor and must remain aligned with image resizing. AGUVIS++ material requires images from upstream AGUVIS and OS-Atlas releases; this repository is not a self-contained copy of every referenced image collection.", "splits": [], "supervision": ["icon grounding", "component grounding", "layout grounding", "refusal"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/Jedi/tree/main", "gated": false, "hub_id": "xlangai/Jedi", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T09:33:18Z", "role": "primary", "sha256": "63ff72bea0a2cb9218093d7a5a420b39c01861aef66141433d7c56c1e0ad05ad", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/xlangai/Jedi"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:38Z", "role": "primary", "sha256": "1d4d15c0cd27837602d55f0fda1ed91d9e8a98b33b3e3c4ba64decae42399558", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/xlangai/Jedi/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T09:33:18Z", "role": "primary", "sha256": "63ff72bea0a2cb9218093d7a5a420b39c01861aef66141433d7c56c1e0ad05ad", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xlangai/Jedi"}], "id": "url:e062cf1445eaee0f3ade", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T09:33:18Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T09:33:18Z", "url": "https://huggingface.co/datasets/xlangai/Jedi"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/Jedi"}], "name": "Jedi", "notes": "The dataset card requires coordinate scaling to stay aligned with the image processor. Follow the publisher recipe before using the grounding labels.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/datasets/xlangai/Jedi", "target_id": "arxiv:2505.13227", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Synthetic GUI grounding data associated with UI decomposition research and the OSWorld-G benchmark.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/xlangai/Jedi"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 train", "qualifier": "exact", "unit": "question–answer examples", "value": 307315}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 train", "qualifier": "exact", "unit": "web pages", "value": 4549}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 dev", "qualifier": "exact", "unit": "question–answer examples", "value": 52826}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 dev", "qualifier": "exact", "unit": "web pages", "value": 913}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 test", "qualifier": "exact", "unit": "question–answer examples", "value": 40357}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 test", "qualifier": "exact", "unit": "web pages", "value": 985}], "evidence_urls": ["https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "https://huggingface.co/api/datasets/X-LANCE/WebSRC_v1.0", "https://x-lance.github.io/WebSRC/"], "formats": ["CSV", "HTML", "JSON", "ZIP packages"], "modalities": ["questions and answers", "HTML", "screenshots", "layout metadata"], "notes": "Each dataset.csv row is a QA example; web pages and QA examples are different units. Use the v1.0 split table rather than merging the older project headline into it. The card retains email-evaluation instructions, while the manifest now includes test answers; no submission was made. Payload archives were not downloaded.", "splits": ["train", "dev", "test"], "supervision": ["extractive or yes/no answers", "answer DOM-element references"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/tree/7aa0bc6efc7ef43f68c192e2091108541acbaf1a", "gated": false, "hub_id": "X-LANCE/WebSRC_v1.0", "hub_revision": "7aa0bc6efc7ef43f68c192e2091108541acbaf1a", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:16Z", "role": "primary", "sha256": "061e5bcaa49e819f0133afb912ad5ce20a8c2b6c28ca60319a4635d6ce835ddb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:29Z", "role": "primary", "sha256": "a7976dd931c3fcf11de2456614f95eb071978599bd70fb89c0ea009d31da4300", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/X-LANCE/WebSRC_v1.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:04Z", "role": "primary", "sha256": "3c2c601d8b4ad2f379cac9b97a2cbb901883d9bc6c1a682359f9b4135e6dd4bc", "supports": ["notes", "details.dataset_profile"], "url": "https://x-lance.github.io/WebSRC/"}], "id": "url:e0ba328ae1a643b0a5aa", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:35:16Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-18T04:35:16Z", "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0"}], "name": "WebSRC v1.0", "notes": "Public release includes train+dev and test ZIP packages plus a test-answer JSON file. The v1.0 card’s exact split table differs from the older project-page headline of 0.44M questions/6.5K pages; counts are kept with their sources. This is static webpage understanding, not an interactive browser-agent trajectory dataset. CC-BY-4.0 is the data-card declaration.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "target_id": "arxiv:2101.09465", "type": "documented_by"}], "review_status": "curated", "scope": "adjacent", "summary": "Webpage segments, questions, answers, HTML, screenshots and layout metadata for structural reading comprehension.", "topics": ["screen-understanding", "grounding"], "updated_at": null, "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md", "https://huggingface.co/api/datasets/CognitiveKernel/WebEvolver"], "formats": ["JSONL"], "modalities": [], "notes": "Two purpose-specific files are released: webevolver_policy_sft.jsonl and webevolver_world_model_sft.jsonl. They are not assumed disjoint train/test splits, and no modality or example count is inferred from filenames alone.", "splits": [], "supervision": ["agent-policy supervised fine-tuning", "web-world-model supervised fine-tuning"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver/tree/13f4f45b359e138a45b1c89bd4e913298c586ca9", "gated": false, "hub_id": "CognitiveKernel/WebEvolver", "hub_revision": "13f4f45b359e138a45b1c89bd4e913298c586ca9", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}, "repository_url": "https://github.com/Tencent/SelfEvolvingAgent/tree/main/WebEvolver"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:17Z", "role": "primary", "sha256": "36ccdce12c1b1512ad4b21326c5d0c0379c73e85f0b8150779d43240b54686de", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:29Z", "role": "primary", "sha256": "cbef3f8037405d311ed9fe8aa620ec3734f91c77ca8f4209fbe98167fb33bee8", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/CognitiveKernel/WebEvolver"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "447dfa1e52789f7f3f111e9e0d453fb1f9fd8bf2d7b858f28fbe21a2239b4b2e", "supports": ["details.repository_url"], "url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/WebEvolver/README.md"}], "id": "url:e126bbd9729ebdaf05ab", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:35:17Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:17Z", "url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver"}], "name": "WebEvolver SFT data", "notes": "The card and public manifest identify separate policy-model and world-model JSONL files. No row total or complete record schema is documented in the inspected card; the files were not downloaded as training samples. MIT is the publisher’s data declaration.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md", "target_id": "arxiv:2504.21024", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Two JSONL collections for supervised fine-tuning the WebEvolver agent policy and web world model.", "topics": ["web-navigation", "world-models", "imitation-learning"], "updated_at": null, "url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/microsoft/OmniParser-v2.0/tree/main", "gated": false, "hub_id": "microsoft/OmniParser-v2.0", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/README.md", "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect/LICENSE", "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect_v3/LICENSE", "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_caption/LICENSE"], "reason": "The top-level Hub metadata says MIT, but the same card and folder licenses distinguish the original detector from the new detector and captioner. Preserve one model-bundle identity with a mixed license declaration."}}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "f4ebaafcc0ab0b9e9cc55cb587678e5c83a5bbd64469fb2cc52bc63843c5cc6d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/microsoft/OmniParser/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "1ea8af0cfd9cd716de8af8b78b2451470e1e1f22233a7286c5f40e1f496dc4be", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/OmniParser-v2.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "6ae154d65bfc2755703440be02acf63a37b741a1101b7c42b666da93a4ed2310", "supports": ["summary", "licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:09:13Z", "role": "primary", "sha256": "4df3c306dddaaf4baffdff5ca820cc679ac8cd6dc263c6a74517783e42fa7a3b", "supports": ["licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:57Z", "role": "primary", "sha256": "646f8936b8ddcd14e13e578ff6857e368780b0d1a4f6066bee89211923a373e2", "supports": ["licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect_v3/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:09:13Z", "role": "primary", "sha256": "d9a1b1e30d633d5732ea18e3cba9538d293ebc53e1a9e4e96ab739e0c5c4f1cb", "supports": ["licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_caption/LICENSE"}], "id": "url:e21ca7c859544dbf5fb6", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher per-folder model license declarations and retained LICENSE files; this is a component map, not one SPDX license for the full bundle.", "declared": "Mixed by component: icon_detect = AGPL-3.0; icon_detect_v3 = MIT; icon_caption = MIT", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/OmniParser-v2.0"}], "name": "OmniParser v2 model bundle", "notes": "The manifest contains icon_detect/model.pt, icon_detect_v3/model.pt and icon_caption/model.safetensors. Licenses differ by folder: original YOLOv8 detector AGPL-3.0; YOLOv9-E detector and Florence-2 captioner MIT. The card’s top-level MIT metadata must not cover the entire bundle. Components remain one resource; no binaries were executed.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "A screen-parsing weight bundle containing icon-region detectors and an icon captioner for converting screenshots into agent-readable interface elements.", "topics": ["screen-understanding", "grounding"], "updated_at": null, "url": "https://huggingface.co/microsoft/OmniParser-v2.0"}
{"aliases": [], "authors": [], "details": {"base_model": "Qwen/Qwen2-VL-7B", "download_url": "https://huggingface.co/osunlp/Dreamer-7B/tree/main", "languages": ["en"]}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:10:45Z", "role": "primary", "sha256": "f15e05cc2718948e344bd1fb943402797ef2e3f2130a9ce2fd94d419210c0574", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/osunlp/Dreamer-7B/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T11:10:45Z", "role": "primary", "sha256": "f15e05cc2718948e344bd1fb943402797ef2e3f2130a9ce2fd94d419210c0574", "supports": ["licenses.weights"], "url": "https://huggingface.co/osunlp/Dreamer-7B/raw/main/README.md"}], "id": "url:e3c1a2a2d556e4728953", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T11:10:45Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-16T11:10:45Z", "url": "https://huggingface.co/osunlp/Dreamer-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/osunlp/Dreamer-7B"}], "name": "Dreamer-7B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/osunlp/Dreamer-7B/raw/main/README.md", "target_id": "arxiv:2411.06559", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "WebDreamer model checkpoint that predicts the consequences of web actions from screenshots and textual descriptions.", "topics": ["world-models", "planning"], "updated_at": null, "url": "https://huggingface.co/osunlp/Dreamer-7B"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL/tree/main", "gated": false, "hub_id": "microsoft/GUI-Actor-7B-Qwen2-VL", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "c76d9811dc6152d1b33bec7f19e808c1882e86cca79bd3fea90075ac9df5c7e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/microsoft/GUI-Actor-7B-Qwen2-VL"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "c76d9811dc6152d1b33bec7f19e808c1882e86cca79bd3fea90075ac9df5c7e6", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/GUI-Actor-7B-Qwen2-VL"}], "id": "url:e3e5dd82a3102689c4ea", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL"}], "name": "GUI-Actor 7B Qwen2-VL", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/microsoft/GUI-Actor-7B-Qwen2-VL", "target_id": "github:microsoft/gui-actor", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "GUI-Actor checkpoint fine-tuned from Qwen2-VL-7B-Instruct, with about 8.4B BF16 parameters, whose config names a pointer variant of the Qwen2-VL class rather than the stock one, under MIT.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL"}
{"aliases": [], "authors": [], "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/osunlp/UGround-V1-7B/tree/main", "gated": false, "hub_id": "osunlp/UGround-V1-7B", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "080c9245276e9ffe5911551a66acd7e91b478f85b96c795d6079465d92580f18", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/osunlp/UGround-V1-7B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "080c9245276e9ffe5911551a66acd7e91b478f85b96c795d6079465d92580f18", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/osunlp/UGround-V1-7B"}], "id": "url:e554295a7f3f5c9ad0d6", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/osunlp/UGround-V1-7B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/osunlp/UGround-V1-7B"}], "name": "UGround V1 7B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/api/models/osunlp/UGround-V1-7B", "target_id": "github:osu-nlp-group/uground", "type": "variant_of"}], "review_status": "curated", "scope": "core", "summary": "UGround-V1 checkpoint fine-tuned from Qwen2-VL-7B, with about 8.3B BF16 parameters in the standard Qwen2-VL architecture, provided as safetensors and a PyTorch file under Apache-2.0.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/osunlp/UGround-V1-7B"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md"], "formats": [], "modalities": ["images", "prompts", "actions", "responses"], "notes": "The card reports a cleaned size of 3.12M without defining the counting unit, so no sample total is inferred. The data card declares noncommercial share-alike terms, separately from the model-weight license.", "splits": [], "supervision": ["web-world-model prediction"]}, "download_url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/tree/main", "modalities": ["image", "text"]}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "e656246f09340fed08136cf41a8b79880e19e1397ddd176410688ddee4e6365c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "e656246f09340fed08136cf41a8b79880e19e1397ddd176410688ddee4e6365c", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md"}], "id": "url:eadc6cba9ea46c38dd1d", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-nc-sa-4.0", "retrieved_at": "2026-10-16T11:09:28Z", "url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data"}], "name": "Dreamer V1 Data", "notes": "The card reports a cleaned size of 3.12M without explicitly defining its counting unit. The catalog leaves the exact sample count unknown. Its data license differs from the model-weight license.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md", "target_id": "arxiv:2411.06559", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Web-world-model training data with prompts, images, action descriptions, and responses, released with the WebDreamer project.", "topics": ["world-models", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "single-turn portion", "qualifier": "approximately", "unit": "QA pairs", "value": 44000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "multi-turn portion", "qualifier": "approximately", "unit": "dialogues", "value": 6000}], "evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"], "formats": ["JSON", "Parquet"], "modalities": ["screenshots", "text", "bounding boxes"], "notes": "Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.", "splits": [], "supervision": ["single-turn question answering", "multi-turn dialogue"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/yiye2023/GUIChat/tree/main", "gated": false, "hub_id": "yiye2023/GUIChat", "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "https://huggingface.co/datasets/yiye2023/GUIChat/raw/main/README.md"], "reason": "Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict."}}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "3ccb4a9ded4f45aa334f0df5dc9a0f8963c645857e8442c46efe1d9cf4a4360b", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/yiye2023/GUIChat"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "4bcf87ecfbbb8e07a01b21415a970c8b53a5283bf6872b657040d3f45c9241f7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yiye2023/GUIChat/raw/main/README.md"}], "id": "url:edf38d43800be8c955ae", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yiye2023/GUIChat"}], "name": "GUIChat", "notes": "Public JSON annotations and Parquet image packages are present. The project README declares CC-BY-4.0 for datasets, while this Hub card declares Apache-2.0; these conflicting declarations are retained without asserting a single confirmed data license. The project code’s MIT declaration is separate.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Conversational GUI supervision with text-rich images and bounding boxes, organized as single-turn questions and multi-turn dialogues.", "topics": ["grounding", "data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/yiye2023/GUIChat"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "69e029d01dda3806f88abfaee01da08c671adcb3555f1314e975ab8942a686ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://browser-use.com/"}], "id": "url:ef244017b1460269fc68", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://browser-use.com/"}], "name": "Browser Use Cloud", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://browser-use.com/", "target_id": "organization:browser-use", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Hosted browser-agent platform from Browser Use.", "topics": ["web-navigation"], "updated_at": null, "url": "https://browser-use.com/"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "474a27cf644ac5f8400a023086fb8d356e6148de71d1851337215f88624852d9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://openai.com/index/introducing-operator/"}], "id": "url:f0584842138a52158f51", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/introducing-operator/"}], "name": "Operator", "notes": "Historical product entry; the official page describes its integration into ChatGPT agent. Do not infer current standalone availability.", "published_at": null, "relations": [{"evidence_url": "https://openai.com/index/introducing-operator/", "target_id": "organization:openai", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Historical OpenAI browser-agent research preview.", "topics": ["web-navigation", "history"], "updated_at": null, "url": "https://openai.com/index/introducing-operator/"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/tree/main", "gated": false, "hub_id": "LZXzju/Qwen2.5-VL-3B-UI-R1"}, "environments": ["web", "desktop", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "620fbee20a6002bbc13562030ce4988d41ef17b0c033ec6872d83741f597adb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "6a206702a1a4faef428460e07f52aaf8fd1f1a144d711d807ca42479b248f697", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/LZXzju/Qwen2.5-VL-3B-UI-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f55ac6a0c50d298d46c5f7083a0879959fd7cd3c1d91496af295ebf2afcc13a3", "supports": ["environments"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "620fbee20a6002bbc13562030ce4988d41ef17b0c033ec6872d83741f597adb8", "supports": ["licenses.weights"], "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md"}], "id": "url:f13474d7aa734a30624d", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1"}], "name": "UI-R1 3B", "notes": "Two safetensors shards and tokenizer/configuration files are publicly ungated. The model card declares MIT; upstream base-model terms are separate. The mutable card compares v1, v2 and UI-R1-E results without pinning every result to a commit, so this record identifies the Hub release and does not import an unversioned score.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md", "target_id": "arxiv:2503.21620", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL checkpoint trained with reinforcement learning for GUI action prediction and evaluated on mobile, desktop and web grounding tasks.", "topics": ["grounding", "reinforcement-learning", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo1-7B/tree/main", "gated": false, "hub_id": "Hcompany/Holo1-7B"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "a7e185380bb02150d761651420cccc9c4a57f67c70a5dec3b7f85af63fb8773b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "66076674ebbdcf9faeb408ebd137ac1283fb56b1011acd7037ebb73e09ba749d", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo1-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "a7e185380bb02150d761651420cccc9c4a57f67c70a5dec3b7f85af63fb8773b", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md"}], "id": "url:f35cfcbaeb79663eabeb", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:44:50Z", "url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo1-7B"}], "name": "Holo1 7B", "notes": "Four public safetensors shards and example navigation/localization files are present. The model card declares Apache-2.0. Reported Surfer-H results depend on the assembled policy/localizer/validator configuration and are not imported as standalone checkpoint scores.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md", "target_id": "arxiv:2506.02865", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-based action model released for the Surfer-H web-agent system, supporting policy, interface-localization and validation roles.", "topics": ["web-navigation", "grounding", "planning"], "updated_at": null, "url": "https://huggingface.co/Hcompany/Holo1-7B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/showlab/ShowUI-pi/tree/main", "gated": false, "hub_id": "showlab/ShowUI-pi"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "64993ab5dc8205e25bda2576055b7e252104c67d4e6eeecb33a1a9742415f67e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/showlab/ShowUI-pi/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "ae12ece91946d79c5c6b578d6db07b5f095822be1ca3bc14bb3dc04a771bb010", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/showlab/ShowUI-pi"}], "id": "url:f70aa7c95824a8729516", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/showlab/ShowUI-pi"}], "name": "ShowUI-π checkpoint", "notes": "A model.safetensors file and policy pre/postprocessor configuration are publicly ungated. The card describes a SmolVLA-based GUI drag-and-drop model despite its generic robotics pipeline tag. No weight-license declaration or specific operating-system coverage is provided.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/showlab/ShowUI-pi/raw/main/README.md", "target_id": "arxiv:2512.24965", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A compact visual-action checkpoint that predicts continuous GUI drag trajectories from a screenshot and a natural-language instruction.", "topics": ["grounding", "action-prediction"], "updated_at": null, "url": "https://huggingface.co/showlab/ShowUI-pi"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "web subset", "qualifier": "approximately", "unit": "instructions", "value": 2900000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "web subset", "qualifier": "approximately", "unit": "images", "value": 173000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "mobile subset", "qualifier": "approximately", "unit": "instructions", "value": 1100000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "mobile subset", "qualifier": "approximately", "unit": "images", "value": 104000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "desktop subset", "qualifier": "approximately", "unit": "instructions", "value": 150000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "desktop subset", "qualifier": "approximately", "unit": "images", "value": 7800}], "evidence_urls": ["https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"], "formats": ["JSON annotations", "image archives"], "modalities": ["screenshots", "instructions", "element captions"], "notes": "The card identifies deprecated desktop files with screenshot-mapping errors and names corrected releases. Counts are rounded publisher figures for separate platform subsets; instructions and images are not interchangeable.", "splits": [], "supervision": ["element grounding", "element descriptions"]}, "download_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/tree/main", "modalities": ["image", "text"]}, "environments": ["web", "android", "desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "8713ca2a6b1962d5dce7da132e1135898d467053dc6b53b4463a0d43a6433bcb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "8713ca2a6b1962d5dce7da132e1135898d467053dc6b53b4463a0d43a6433bcb", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"}], "id": "url:f8566bd0394d36ed4f8e", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T10:42:13Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-16T10:42:13Z", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data"}], "name": "Aria-UI Data", "notes": "The publisher flags original desktop files as deprecated because of screenshot-mapping issues, and identifies corrected versions in the dataset card.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "target_id": "arxiv:2412.16256", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "GUI element descriptions and grounding instructions spanning web, mobile, and Ubuntu desktop interfaces.", "topics": ["grounding", "screen-understanding"], "updated_at": null, "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data"}
{"aliases": ["url:a6db151f939a6748ff76", "showlab/ShowUI-desktop", "https://huggingface.co/datasets/showlab/ShowUI-desktop"], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/raw/main/README.md", "population": "train split declared in the dataset card", "qualifier": "exact", "unit": "examples", "value": 7496}], "evidence_urls": ["https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/raw/main/README.md"], "formats": ["Hub table with an image column"], "modalities": ["screenshots", "text", "bounding boxes", "points"], "notes": "The 8K release name is not an exact record count; the card declares 7,496 examples. Images and original annotations come from OmniAct and are augmented with generated descriptions. Examples are not asserted to be unique screenshots.", "splits": ["train"], "supervision": ["instruction-to-element grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/tree/main", "gated": false, "hub_id": "showlab/ShowUI-desktop", "languages": [], "modalities": ["image", "text", "timeseries"], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "84bd2ef05ea837eaeebb248002b2942dd5c72f072beb29a9bd70eda1ea433a31", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-desktop-8K"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:38Z", "role": "primary", "sha256": "577a78efa03cde05141c23033222014094681cb5d04786dbb4a698df8543b3df", "supports": ["details.dataset_profile", "relations"], "url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "84bd2ef05ea837eaeebb248002b2942dd5c72f072beb29a9bd70eda1ea433a31", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-desktop-8K"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:57Z", "role": "primary", "sha256": "57eaa720c20e9e3a4428c8292c150e1949f3372c75c6890e424ba3c6600a5e5d", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-desktop"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:58Z", "role": "primary", "sha256": "57eaa720c20e9e3a4428c8292c150e1949f3372c75c6890e424ba3c6600a5e5d", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-desktop-8K"}], "id": "url:f976dec3d1c5e0bedbd7", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K"}], "name": "ShowUI Desktop", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/raw/main/README.md", "target_id": "github:showlab/showui", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "7,496 desktop grounding examples built from OmniAct screenshots, with annotations rewritten by GPT-4o to describe element appearance, spatial relations and function.", "topics": ["data-generation"], "updated_at": null, "url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/raw/main/README.md", "population": "Hub split named test", "qualifier": "exact", "unit": "rows", "value": 3630}], "evidence_urls": ["https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/raw/main/README.md", "https://huggingface.co/api/collections/LangAGI-Lab/web-shepherd-advancing-prms-for-reinforcing-web-agents-682b4f4ad607fc27c4dc49e8"], "formats": ["Parquet"], "modalities": ["screenshots", "text", "action history"], "notes": "This is the checklist-generation release rather than the separate preference-pair collection. No unique-task count or data permission is inferred.", "splits": ["test"], "supervision": ["checklist generation", "subgoal analysis"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/tree/main", "gated": false, "hub_id": "LangAGI-Lab/WebPRMCollection_checklist_generation", "languages": [], "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "2e4142675f2cb994c21a7620a5188b10147ab03326a2047946886b5916789ec2", "supports": ["name", "kind", "identity", "release_status", "details", "summary"], "url": "https://huggingface.co/api/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "3693cb307065a6ced63a4789a7f592b5eba0ad6a5ffe1a3caf0649decd5cb6e8", "supports": ["details"], "url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:05Z", "role": "primary", "sha256": "647d9e7bdad5fe2c03a8cfcd34d93f3b2bf7c1553b517a4a843bc0113fa47456", "supports": ["identity", "summary"], "url": "https://huggingface.co/api/collections/LangAGI-Lab/web-shepherd-advancing-prms-for-reinforcing-web-agents-682b4f4ad607fc27c4dc49e8"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "07431058d8b9a71b80e75b7a74a40a8346c1503eec240628459cc4d4e2f0802f", "supports": ["summary", "environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "2e4142675f2cb994c21a7620a5188b10147ab03326a2047946886b5916789ec2", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation"}], "id": "url:fcc98b149272bbf31b05", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation"}], "name": "WebPRM checklist-generation data", "notes": "Five public Parquet shards are present. This collection provides checklist-generation supervision and is distinct from the previously cataloged chosen/rejected preference data. The Hub calls its split test; that label is preserved. No data license is declared.", "published_at": null, "relations": [], "review_status": "curated", "scope": "core", "summary": "Web-task context, starting observations and action history paired with structured checklists and subgoal analyses for checklist generation.", "topics": ["reward-models", "data-generation", "web-navigation"], "updated_at": null, "url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation"}
{"aliases": [], "authors": [], "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "6232364481bbae45a3fcec1d1aca858493baff0693ececa98b09b610183e6796", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://e2b.dev/"}], "id": "url:fda52d641c6b482bda7b", "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://e2b.dev/"}], "name": "E2B", "notes": "", "published_at": null, "relations": [{"evidence_url": "https://e2b.dev/", "target_id": "organization:e2b", "type": "published_by"}], "review_status": "curated", "scope": "core", "summary": "Sandbox infrastructure for executing agent workloads.", "topics": ["sandbox"], "updated_at": null, "url": "https://e2b.dev/"}
{"aliases": [], "authors": [], "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/h-siyuan/ScreenDrag/raw/main/README.md"], "formats": ["Parquet", "episode files"], "modalities": ["screen images", "instructions", "previous state", "continuous actions"], "notes": "Raw and Parquet folders are alternative views of the same data. Each sample describes a dragging task over multiple timesteps; action vectors encode button state and coordinates.", "splits": ["train", "test"], "supervision": ["continuous GUI dragging"]}, "download_url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag/tree/main", "modalities": ["image", "text", "continuous-actions"]}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:12:56Z", "role": "primary", "sha256": "39c7c028f7ef5ca5c75f79b59a107c7e25444776095a2a08bb7535c4f523f7fa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "88df552ccbe80efd7020905d77d526f4f9efe2d09a11fed2714b083988f9dcef", "supports": ["relations"], "url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md"}], "id": "url:fe43af1801032be708cc", "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:12:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag"}], "name": "ScreenDrag", "notes": "Raw and Parquet folders are alternative views of the same data and should not be counted as independent samples. The inspected card did not state a license.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md", "target_id": "github:showlab/showui-pi", "type": "component_of"}], "review_status": "curated", "scope": "core", "summary": "GUI dragging data with screen observations, prior state, and continuous pointer actions, provided as trajectory files and Parquet representations.", "topics": ["grounding", "evaluation"], "updated_at": null, "url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag"}
{"aliases": [], "authors": [], "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-2B", "gated": false, "hub_revision": "21e684b39c33e7efaf055f000374d74f37556954", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "08d14b9dc3010365e543cbb6707b67c5d792118451bbc2d0b70f1f0280cb20e1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "7b3242c3943537b6a26d8650a78c45044dce06a031ca2ce8bc67d5e482a332ce", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-2B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "08d14b9dc3010365e543cbb6707b67c5d792118451bbc2d0b70f1f0280cb20e1", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md"}], "id": "url:fec1c5aa8fd93cb78923", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B"}], "name": "Seg2Click Qwen3.5 2B", "notes": "The public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.", "published_at": null, "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "Seg2Click checkpoint using a 2B Qwen3.5 backbone to predict a mask and select an interface click.", "topics": ["grounding"], "updated_at": null, "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B"}
{"aliases": [], "authors": [], "details": {"disabled": false, "download_url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/tree/main", "gated": false, "hub_id": "OpenGVLab/ScaleCUA-7B"}, "environments": ["web", "desktop", "android", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "e45116466b8b9ef3e93992a1efca8fa5b921a8b041297933210c9f552767ffb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:19Z", "role": "primary", "sha256": "bad2617fd64633212a30b78d99fb053a662378ab50b42722a04af4104ac2248f", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/OpenGVLab/ScaleCUA-7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:54:55Z", "role": "primary", "sha256": "0538701c860afedd5d63c3839039699d8e5a14977d5c339b187ed13ad666c8fa", "supports": ["environments", "relations"], "url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "e45116466b8b9ef3e93992a1efca8fa5b921a8b041297933210c9f552767ffb8", "supports": ["licenses.weights"], "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/raw/main/README.md"}], "id": "url:ff7b1873a3eead89607a", "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:45:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:45:20Z", "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B"}], "name": "ScaleCUA 7B", "notes": "Four public safetensors shards and model/tokenizer configuration are present. The card declares Apache-2.0. The model card’s Paper link points to the code repository, so the arXiv association uses the repository’s explicit citation rather than treating that link as a paper URL.", "published_at": null, "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/README.md", "target_id": "arxiv:2509.15221", "type": "documented_by"}], "review_status": "curated", "scope": "core", "summary": "A Qwen2.5-VL-based computer-use checkpoint with direct grounding and agent action modes across desktop, mobile and web interfaces.", "topics": ["grounding", "action-prediction", "planning"], "updated_at": null, "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B"}